Skip to main content

Full text of "Digitization In The Real World"

See other formats



Metropolitan 


m 


New York 




Library Council 




Abraham Lincoln Brigade Archives, New York University 
American Museum of Natural History 
Barnard College 
Center for Jewish History 

Central/Western Massachusetts Automated Resource Sharing System 
Central Massachusetts Regional Library System 
Chelsea District Library, Michigan 
Clemson University 
Colorado State University 
Columbia University 
County of Brant Public Library 
Fort Hays State University 
Illinois State University 
Jewish Theological Seminary 
Kansas Cosmosphere and Space Center 
Leo Baeck Institute 
Louisiana State University 
Louisiana State University Health Sciences Center 
Massachusetts Regional Library System 
Metropolitan New York Library Council 
Middle Tennessee State University 
Museum of Seminole County History in Sanford, Florida 
New York Public Library 
Orange County Library System, Florida 
Orange County Regional History Center, Florida 
Oregon State University 
Pratt Institute 

Queens College, City University of New York 
Rollins College 
Rutgers University 

Southeastern New York Library Resources Council 
Southern Illinois University Edwardsville 
Southern New Hampshire University 
University of Alabama 
University of Central Florida 
University of Iowa ■ 
University of Massachusetts Bost 
Western Carolina University 
Western Massachusetts Regional Library System 
Wisconsin Historical Society 
Yale University f (¥) 



9' 780615" 379982 





m THE 



Lessons Learned from 
Small and Medium-Sized 
Digitization Projects 





Kwong Bor Ng 
& Jason Kucsma 



Digitization in the 
Real World 

Lessons Learned from 
Small and Medium-Sized 
Digitization Projects 



Edited by 

Kwong Bor Ng & Jason Kucsma 



^8 Metropolitan New York Library Council 



i 



Published in the United States of America by 

Metropolitan New York Library Council 

57 East nth Street, 4th floor 

New York, NY 10003-4605 

p: (212) 228-2320 f: (212) 228-2598 

Web site: http://www.metro.org 
ISBN: 978-0-615-37998-2 

Cover Design: Jason Kucsma (illustration by Smartone Design, 
licensed via iStockphoto.com) 

Reviewers Committee: Mark F. Anderson, Jill Annitto, Anna Craft, Jody 
DeRidder, Renate Evers, Wei Fang, Maureen M. Knapp, Sue Kunda, 
Mandy Mastrovita, Ken Middleton, Emily Pfotenhauer, Mark Phillipson, 
Alice Piatt, Mary Z. Rose, Stacy Schiff, Jennifer Weintraub, Andrew 
Weiss. 

Copyright © 2010 by Metropolitan New York Library Council. No part of 
this book may be reproduced in any form or by any means, electronic or 
mechanical, including photocopying, without written permission from 
the publisher. 

The views expressed in this book are those of the authors, but not 
necessarily those of the publisher. 



ii 



About the Editors 



Kwong Bor Ng (kbng@qc.cuny.edu) is an associate professor at the 
Graduate School of Library and Information Studies, Queens College, 
CUNY. His research interests are in the technical and technological 
areas of knowledge representation, organization, storage and retrieval. 
Dr. Ng has published more than 20 academic papers in refereed 
journals. His most recent publications include Using XML: A How-to- 
do-it Manual and CD-ROM for Librarians (published by Neal 
Schuman in 2007) and Collaborative Technologies and Applications 
for Interactive Information Design: Emerging Trends in User 
Experiences (co-edited with S. Rummler, published by Information 
Science Reference, IGI-Global, Inc. in 2009.) 

Jason Kucsma 0kucsma@metro.org) is the Emerging Technologies 
Manager at the Metropolitan New York Library Council where he 
manages METRO'S Digitization Grant Program and is the point 
person for member inquiries related to the resources, training and 
referral services associated with digitization, digital preservation and 
emerging technologies issues. Jason received his M.A. in American 
Culture Studies from Bowling Green State University and an M.L.S. 
from the University of Arizona School of Information Resources and 
Library Science. He is currently a part-time lecturer in Rutgers's 
Library and Information Science graduate program, and is a recent 
graduate of ALA's 2009 Emerging Leaders Program. 



iii 



Table of Contents 



Foreword vii 

Preface ix 

Part I — Small is Beautiful: Planning and Implementing 
Digitization Projects with Limited Resources 

DIY Digitization: Creating a Small-scale Digital Zine 

Exhibit l 

Digitizing Civil Rights: An Omeka-based Pilot Digital 

Presence for the Queens College Civil Rights Archive 23 

Digitization on a Dime: How a Small Library and a Big 
Team of Volunteers Digitized 15,000 Obituaries in Just 
Over a Year 41 

Building the ALBA Digital Library 57 

Digitization and Access of Louisiana Oral Histories: One 

Oral History Center's Experience in the Digital Realm 68 

Digitizing a Newspaper Clippings Collection: a Case Study 

and Framework for Small-Scale Digital Projects 86 

METRO Grant Success Story: Waterways of New York 

Project 100 

Part II - A Diverse Digital Landscape: Digital Collections in 
Public Libraries, Museums, Cultural Heritage Institutions, 
and Knowledge-Based Organizations 

Managing Rights in a Medium Scale Audio Digitization 

Project 125 

The In-House Digital Laboratory: Possibilities and 

Responsibilities 136 

iv 



Building a Virtual Library - A Case Study at The Library of 

The Jewish Theological Seminary 154 

Local Objects, Local People, Local History: Creating the 

Wisconsin Decorative Arts Database 172 

Digitizing the Rare Book Collection of the Leo Baeck 

Institute 185 

The Caprons of Paris: A Digitization Project in a Small 

Library System 195 

The Mass. Memories Road Show: A State-Wide Scanning 

Project 223 

Picturing the Museum: Education and Exhibition at the 

American Museum of Natural History 242 

Part III - The Digital Campus: Digitization in Universities 
and Their Libraries 

Developing an Institutional Repository at Southern New 

Hampshire University: Year One 261 

Digitization of the Yale Daily News Historical Archive 274 

New Jersey Digital Legal Library 289 

Pratt Institute: A Historical Snapshot of Campus and Area 302 

Scaling Back for an "Experimental" Collection 321 

From Confusion and Chaos to Clarity and Hope: 

Reorganization of Work Flows, Processes, and Delivery 

for Digital Libraries 333 

Digitizing Colorado State University's Historic Photograph 

Collection: A Case Study 355 

Entering the Digitization Universe: One Catalog Librarian's 

Experience at an Academic Library 370 



V 



From Argentina to Zambia: Capturing the Digital A to Z's 

of a Child Art Collection 383 

Special Collections, Digitization, and the Classroom: A 

New Model 402 

Part IV - One Plus One is Greater Than Two: Collaborative 
Projects 

Digital Treasures: The Evolution of a Digital Repository in 

Massachusetts 423 

Collaborative Digitization Goes Local 435 

Picturing the Past and Planning for the Future: Central 

Florida Memory 450 

Apollo 13.0: Digitizing Astronaut Jack Swigert's Apollo 

Documents 470 

Collaborative-Centered Digital Curation: A Case Study at 

Clemson University Libraries 490 

The Craft Revival Project: Library Leadership in Creating 

Connections between Small Cultural Institutions 503 

Hudson River Valley Heritage: A Journey in Collaborative 

Digitization 518 

Collaborating for Success: A Cross-Departmental 

Digitization Project 541 

Using Omeka to Build Digital Collections: The METRO 

Case Study 556 

vi 



Foreword 



Dottie Hiebing (METRO Executive Director) 

For more than 45 years, METRO has worked to provide opportunities 
for libraries to share best practice strategies to address many critical 
needs. In these efforts, we have often seen that the best learning 
comes through examples of libraries that have addressed important 
challenges successfully - and sometimes not so successfully. This has 
been especially true in efforts to support large and small digitization 
projects. 

For more than a decade - and continuing today - digitization 
has been established as an essential focus for many libraries as well as 
for research centers, museums, and cultural and arts organizations. 
METRO has worked to support our members in these efforts with a 
range of grants, training programs and instructional materials. 

Digitization in the Real World represents a significant new 
milestone in our commitment to providing library professionals with 
the hands-on experience and guidance they need to plan, execute, and 
manage digitization projects over the long term. In many ways, the 
examples presented in this volume show library professionals how to 
maximize the value and impact of digitization efforts for their libraries 
and their users. 

This book also represents the first self-published text METRO 
has ever sponsored. As we continually look for new ways to help 
libraries stay ahead of the curve in digitization, technology and other 
areas, this strategy has the clear potential to be a major focus of our 
work in the years ahead. We will welcome your feedback and look 
forward to seeing how self-published materials such as this can 
support our mission and your needs moving forward. 



vii 



On behalf of METRO, I would like to congratulate and thank 
editors Kwong Bor Ng and Jason Kucsma and all of the members of 
the library community who supported this project and who 
contributed of their time and insight in the development of these 
outstanding digitization case studies. They have created a vital new 
resource to help libraries continue to advance important digitization 
projects, and their efforts will have a profound and lasting impact on 
the future of these efforts in the years ahead. 



viii 



Preface 



Kwong Bor Ng & Jason Kucsma 

For more than a decade, digitization has been both a critical need and 
a formidable challenge for libraries, archives, and museums around 
the world. To support these important projects, the Metropolitan New 
York Library Council (METRO) has been awarding annual grants to 
support digitization projects in New York City and Westchester 
County since 2005. Thus far, METRO has provided support for 
approximately 40 digitization projects at 25 different institutions. In 
those five years, we have learned a great deal about managing 
digitization projects effectively. In these efforts, METRO members 
have also shared best practice strategies in digitization through 
project showcase events and through the work of the METRO- 
sponsored Digitization Special Interest Group. 

All digitization projects begin with some critical questions. How 
do we start a digitization project? What standards should we use for 
digital conversion and metadata? What are the best practices for 
workflow? What equipment or software should we use? Should we 
digitize in-house or outsource? What organizational or technological 
obstacles should we anticipate, and how should we negotiate them? 
Where can we turn for help in the middle of a project? 

Naturally, the response to these questions will differ for different 
institutions. Even discrete projects within an institution will have 
many unique characteristics and challenges. But shared stories of 
successes (and failures) can be immensely helpful in supporting future 
digitization projects. To that end, Professor Ng came to METRO in the 
summer of 2009 with a great suggestion. Why not collect some of the 
most compelling examples of recent digitization projects? Many of us 
are familiar with the large-scale mass digitization projects of recent 
years. But Ng suggested — and we agreed — that there was a great 



ix 



opportunity to share insights from lesser-known examples from the 
"real world." That's not to say that large-scale projects don't pose their 
own unique issues and learning opportunities for librarians, 
archivists, and technologists. But many libraries are more likely to 
proceed with smaller-scale digitization projects made possible by a 
special need or unique opportunity, a first-time grant, or the special 
dedication of a team of library professionals. Collectively, these efforts 
can provide many invaluable perspectives and procedural models. 

This book was initially conceived as an opportunity to highlight 
digitization efforts in the New York metropolitan area. Our research 
quickly showed that there were many other project examples worth 
sharing. The response to our initial call for proposals was 
overwhelming; we received hundreds of chapter proposals from all 
over the world in just the first few months. Contacts from many of the 
world's leading knowledge-based organizations, cultural institutions 
and university libraries presented examples of projects representing a 
wide range of topics, perspectives, approaches, concerns, and lessons- 
learned. 

The effort to choose from among these examples the examples 
that would be presented in the book was a daunting task. We were 
unable to include many great case studies. Each of the chapters 
presented was reviewed in a double-blind peer-review process to 
assess quality, accuracy and relevance. The 34 papers presented in 
this book represent our best effort to present a diverse and 
comprehensive overview of key issues in the management and 
realization of digitization projects. 

We have divided the case studies into four primary groups. The 
first section focuses on small projects. They are digitization endeavors 
that moved forward with limited resources and staffing. The second 
group showcases digitization projects from diverse cultural 
institutions including public libraries, museums, research institutes, 
and cultural organizations. The third group consists of digitization 
projects based on medium-sized collections at universities and their 
libraries. The last group features projects that brought together 



X 



multiple institutions to work in collaboration on a project of mutual 
interest. 

This book would not have been possible without the participation 
and hard work of all of the authors and reviewers involved, including 
those who submitted chapters that we were not able to accommodate. 
We're also greatly indebted to Dottie Hiebing, Executive Director of 
METRO, for recognizing the need for this important resource and for 
supporting this effort from inception. This is the first of what we hope 
will become a series of instructional self-publishing projects 
supported by METRO in the years ahead. 

This is, above all, a book written by practitioners for practitioners 
who together recognize the critical needs and goals in digitization in 
our industry. Our hope is that it will be useful to students who are 
preparing for a career in library or research science and to 
practitioners who will shape the future of digitization for the library 
community. We know reading these stories has been enlightening for 
both of us, and we hope it will be for you as well. Thank you for 
reading. 



xi 



xii 



Part I - Small is Beautiful: 

Planning and Implementing 
Digitization Projects with 
Limited Resources 



DIY Digitization: Creating a Small-scale 
Digital Zine Exhibit 



Melissa L. Jones (College Summit) 



Abstract 

The Barnard Library Zine Collection is an innovative special collection 
of dynamic popular culture artifacts. The zines in the collection 
provide a democratic and vibrant glimpse into the movements and 
trends in recent feminist thought through the personal work of artists, 
writers, and activists. The author finds that in order to improve access 
to and generate interest in such niche collections, institutions have a 
responsibility to overcome barriers to digitization and begin sharing 
their collections online. This chapter discusses the development of 
Barnard's first zine digitization project: the Elections and Protests: 
Zines from the Barnard Library Collection Online Exhibit, launched 
in the summer of 2008. The successful project demonstrates that it is 
possible to build effective and engaging small-scale digital collections 
using simple and inexpensive technologies. 

Keywords: Barnard College Library, Copyleft, Copyright, Education, 
Elections, Lesson plans, Online exhibit, Political zines, Primary 
sources, Protest, Special collections, Zines. 



Introduction 

The Barnard College Library began collecting zines in 2003 in an 
effort to document third wave feminism and riot grrrl culture. Zines 
are self-published, usually inexpensively produced works by writers 
who subscribe to a Do It Yourself (DIY) philosophy. Generally, zines 



1 



2 DIY Digitization 



are created out of an interest to communicate or express ideas that 
might not otherwise find acceptance in the mainstream media. 
Although zines as we know them today were born from the punk 
movement of the early 1970s (Duncombe, 1997, p. 21), they are part of 
a long history of small-run and "amateur" publication. Whether 
calling colonialists to arms in the days of the American Revolution or 
subverting censorship and challenges to free speech in Soviet Russia 
(Wright, 1997), alternative publications are a natural and important 
tool for preserving free speech. 

Although zines are low rent ephemera, several public and 
academic libraries across the country have begun to recognize their 
value. At the forefront of the field, Barnard's collection has nearly 
2,500 holdings providing unmediated access to the voices of young 
women on such subjects as race, gender, sexuality, childbirth, 
motherhood and politics. Zine Librarian Jenna Freedman's outreach 
and advocacy work helps to legitimize zines, not as radical historical 
footnotes but as valid literary and historic works worthy of collection, 
preservation and study. 

As the Zine Intern in summer 2008, my role was to help 
Freedman to increase access to and interest in the Zine Collection. The 
result of my work was Barnard Library's first digital collection, an 
online exhibit entitled, Elections and Protests: Zines from the 
Barnard Library Collection. This project employed a DIY approach to 
digitization, making use of materials and resources at hand to solve 
problems and overcome challenges rather than relying on mainstream 
or out-of-the-box technologies. This project demonstrates that small- 
scale digitization projects can be topical, useful and impactful for a 
variety of stakeholders. 

Literature Review and Needs Assessment 

The literature surrounding zines reveals that, as unique primary 
source documents, they can serve as valuable research tools. 
Alternative press advocates such as librarians Chris Dodge and Jim 
Danky argue that self-published ephemera like zines, handbills, and 
military newspapers can provide a glimpse into a part of history that 



DIY Digitization 3 



includes the voices of marginalized individuals and groups which 
would otherwise be lost were they not collected (Dodge, 2008). 

Dempsey (2006) notes that to collect the ephemeral and radical 
"long tail" is not enough; institutions have a responsibility to provide 
users with access points and contextual materials in order to 
maximize use. Liu (2007) notes that in order to better serve users, 
"academic library Web sites should ... switch the focus from 
presenting information arranged according to library functions and 
resources to providing targeted and customizable tools and services to 
library users ... and give users opportunities to express, share, and 
learn." In addition to their value as historical documents, zines also 
serve as powerful teaching tools for media literacy (Wan 1999; 
Congdon, 2003; Daly, 2005), but scholars and teachers need both 
access to zines and support for teaching with these unique documents 
in order to capitalize on this potential. 

Lesk (2007) acknowledges the legal and philosophical issues that 
are inherent in digitization work, but advocates strongly for 
institutions and copyright holders to work together to overcome 
challenges due to the potential value of digital materials for research. 
In order to support online research, some public and academic 
institutions have begun digitizing their special collections. 
Unfortunately at the time of this project, no public or academic 
institution had moved to digitize their zine collections. 

The lack of high-quality materials for studying and teaching zines 
online makes interacting with the genre impossible for anyone 
without physical access to a collection. Most public and academic 
institutions allow access to their zine collections mainly through 
catalog search. Some institutions occasionally mount online exhibits 
that include scans of zine covers only. 

This has been due, in part, to the same barriers that hinder other 
digitization projects such as prohibitive cost, lack of time, and 
technological limitations. Additionally, zine librarians and scholars 
identify the intrinsically physical nature of the genre as another 
reason not to prioritize zine digitization. Migrating zine content to a 
digital form is seen by many in academia to undermine the very heart 



4 DIY Digitization 



of the genre, which is to be rooted in physical interaction between 
zinester, zine, and reader. Duke University's Zine Librarian argues 
that, "...zines are created by hand, crafted with paper, scissors, tape, 
glue, staples. They were meant to be handed from person to person, 
physically shared. The experience of handling zines in person, turning 
each page to reveal intimate secrets, funny comics, and poetry, can't 
be duplicated on-line. You would get the content, but miss out on the 
physical experience (Wooten, 2009)." Any academic digital Zine 
Collection would need to be very conscious of its treatment of digital 
surrogates. 

Concerns about copyright, permission and privacy create another 
barrier to digitization. Copyright is a sticky issue when it comes to 
zines as a genre, which, by definition are created to be shared. Thus, 
many zines contain a copyleft statement, or some other notation of 
whether the owner has given permission for its contents to be 
reproduced. "Copyleft" is a term coined by open-source software 
pioneers to describe a "flipping" of traditional copyright laws that 
allows content owners to grant broader permission for their work to 
be shared. This "General Public License" can be applied in any 
situation where copyright might apply, including software, books, 
images and music (Soderberg, 2002). Generally, copyleft permission 
or GPL is considered to be conditional; zinesters who select copyleft 
status for their work, or those who claim no legal protection at all, still 
expect to be credited, or at least respected, for their work. It is poor 
zine etiquette to steal, borrow, or sell someone else's zine for personal 
gain. 

Private zine online library and archive groups, run by zinesters 
and fans, have developed to fill the void of zines on the web. The sites 
digitize a large number of zines and serve as valuable repositories of 
content for experts in the field. Because they have grown organically 
from the zine community, these sites maximize their relationships to 
avoid and address concerns about copyright. 

For Barnard Library, the the benefits of digitization digitization 
provided an incentive to overcome potential barriers, challenges and 
costs. An Access and Use Survey of known users administered in 



DIY Digitization 5 



2008 revealed that, while the Zine Collection has a strong contingent 
of feminist and zinester stakeholders, Barnard Library could be doing 
more to attract users outside the immediate scholarly and cultural 
community (see Figure ZINE-i.). 

Answer Options § 




Figure ZINE-1: Results of the Barnard Library Zine Collection Access and Use 
Survey, administered to 25 known users in July 2008. 



6 DIY Digitization 



The survey confirmed that a small-scale digitization project would 
be a valuable addition to Barnard Library's existing services. 81% of 
known users reported that they would use curated, online exhibits 
about zines and zine history. Additionally, 81% of users said that they 
would use digital scans of selected zines. 

The findings of the survey reflect the expectation by users that a 
library's website provide more than just access to information. By 
digitizing the popular and well-respected zine collection, Barnard 
Library could capitalize on the strength of its special collection to meet 
the needs of existing users, attract new users, and fill a need in the 
existing digital zine landscape. Additionally, a digital Zine Collection 
could help spread the word about the value of zines as historical 
documents and teaching tools to a new generation of potential 
stakeholders. 

Project Planning 

After making the decision to create a small digital zine collection, I 
created a project plan that included setting clear goals for the project. 

Goal Setting 

Digitizing even a small portion of Barnard's Zine Collection would 
have many benefits for the institution, its users, and the historical 
record. These included: 

Improving access: Currently, membership in the 
Barnard/Columbia learning community is required in order to secure 
free access to the zine collection. Digitization would allow zines to be 
downloaded and shared easily, improving the ability of people from 
across the globe to access and learn from the collection. 

Raising awareness about zines as legitimate historical 
objects: Freedman's work as an advocate for zine and other radical 
special collections would be complemented by a well-selected 
digitization project that is supported by descriptive and educational 
materials. 

Highlighting Barnard's women's studies collection and 
drawing researchers to the institution: Barnard's Zine 



DIY Digitization 7 



Collection sets its women's studies research collection apart from 
other academic institutions. An online exhibit of materials from the 
Zine Collection could emphasize its uniqueness and eventually bring 
more researchers to the collection. 

Preservation of the collection: Because most zines are 
produced cheaply using poor quality paper and inks, long-term 
conservation can be an issue. Digitizing zines makes their content 
available to future generations of researchers, students, and other 
stakeholders while preserving their physical form. 

Project Scope 

The scale of this project was by necessity very small. No fund was 
designated for the project. The site would need to be built and 
function within the existing Library website's structure; no money was 
available for purchasing a Content Management System or developing 
a complex metadata or image database. All work would need to be 
done using hardware and software already in Barnard Library's 
possession, or available open-source on the Web. As the Zine Intern, I 
would be the sole staff member available to work on the project. 
Freedman would supervise and approve my work. The project would 
need to be completed over the course of my summer internship, 
lasting only 100 hours over the course of 10 weeks. 

Content Selection 

Digitizing the collection as a whole proved to be too time 
consuming and technologically complex given these limitations. 
Selecting a small group of zines in a given theme or subject area to 
digitize first seemed a good model to begin with. Prioritizing 
digitization by demonstrated user need is a model that has been 
successful for other institutions. The University of Warwick in the UK, 
for example, developed an innovative research project in which 
students created digital surrogates of the 18th century French plays 
they used in their coursework (Astbury, 2006). Following a similar 
needs-based model would ensure that Barnard's first digital collection 
would be used by its most immediate stakeholders. 



8 DIY Digitization 

In 2008, the country was gearing up for an historic presidential 
election. Earlier that year, the previous Zine Intern, Julie Turley had 
created an exhibit of "Election and protest themed zines" to connect 
the institution's holdings with current events. The physical exhibit, 
which lived in the library's lobby, featured copies of selected zines and 
photocopied extracts of pages. From the Republican National 
Convention to the presidential election, from deciding to take your 
child to a political rally to challenging politicians to be responsible to 
their electorate, the featured zines addressed participation in — or 
protest against — the American political process. 

The exhibit was a natural fit for this digitization project. The 
presidential election was only months away, we knew conversations 
about the political process would be a hot topic on campus. The 
selected zines offered a little-seen counterpoint to mainstream 
political coverage, rejecting voting as the sole means to make change 
in this nation. Moreover, educators across the nation would be looking 
for ways to talk about elections and the political process in their 
classrooms. It would be an excellent opportunity to demonstrate that 
zines can be relevant political and educational tools. Since zines are 
political in nature and often overtly political in topic, our digital 
collection would be reflective of the genre as a whole, even though we 
could only digitize a small number of zines. Finally, because the zines 
in this subset were already on display in the lobby, we knew that none 
were in need of conservation work or otherwise in danger of being 
damaged by the process of digitization. 

Project Implementation 

To maximize the benefits of digitization while addressing the barriers 
faced by the institution, I undertook a multi-step process for digitizing 
and presenting Barnard zines online. The process, like the zines 
themselves, was low-rent, low-tech, and outside the mainstream. The 
DIY approach was limiting in many ways, but also served as an 
excellent opportunity for learning and innovation. 



DIY Digitization 9 



DIY Digitization Project Timeline - May 5 th through July 24 th , 2008 


1 dol\ 


1 IIIICIIIIC 


Needs assessment and literature review 


Prior to project start 


Goal setting and scope definition 


Prior to project start 


Content selection 


May o 


OtJlltJldtiriy rTlcldUdld 


Maw 1 9 th 
iviay iz 


L'Ompeiiuve lanuscape analysis 


Maw 19 th thrnnnh Mai/ 90 th 

iviay \c. inrougn iviay 


oreanny sue rndps diiu wireiTdrnes 


Maw 19 th fhmiinh Maw 1 ^ th 

iviay iz inrouyn ividy lo 


Convriaht reauests to Dubtish sent to 7inesters 


May 1 9 th 


Scanning and digitization 


May 19 th through June 17 th 


Designing an intuitive user interface 


May 20 th through July 21 st 


Writing original content 


June 1 st through July 18 th 


Usability testing 


July 19 th through July 21 st 


Site launch and publicizing 


July 23 rd through July 24 th 


Evaluation and reporting 


Ongoing 



Figure ZINE-2: Project Timeline - May 5th through July 24th, 2008 



Copyright Status and Securing Permissions 

After selecting the zines to be digitized, securing permission to 
present their content on the web was the next step. Educational use, 
such as the creation of an exhibit, would likely fall within any 
zinester's definition of copyleft. Only one zine of the ten selected, 
"Radical Cheerbook," contained an explicit copyleft statement. We felt 
confident that we could use its content in the exhibition. 

Because the other nine zines selected for this exhibit contained 
some kind of copyright statement or did not contain an explicit 
copyleft statement, an effort was made to contact and secure 
permissions from the original author. This effort was difficult, 
however, since many zines were published using pseudonyms or 
contain contact information that is out of date. To track down the 
zinesters, I used a combination of Google searches, MySpace, and a 
pre-catalog Microsoft Access database that Freedman maintains to 
identify current email addresses. For one zinester, I was only able to 
identify a mailing address, so I sent a letter and awaited a response. 



io DIY Digitization 



By the time the site was ready to go live in mid-July, I received 
written permission to publish from six zinesters, with most expressing 
excitement about the project. One zinester requested that I send scans 
of the specific pages I'd hoped to include before giving permission. At 
the bottom of each zine's page on the site, I made a note that the 
copyright holder had given permission for Barnard to use scans from 
the zines. 

In three cases, I was not able to secure permission before the 
launch of the website. In these cases, I added a note to each zine's 
page that we had made a diligent effort to contact the copyright holder 
and would remove the images used in the event that there was an 
objection. I also made the decision to include only minimal excerpts 
from these zines as compared to the more extensive scans used from 
the zines for which we had permission. 

Site Design & Comparative Landscape Analysis 

Once permissions requests were sent, I focused my work on 
designing the site's architecture and layout. Close analysis of the 
features of similar sites can be a good way to begin planning. In order 
to understand how zines and zine-like publications can be presented 
online, I analyzed five sites with similar collections to Barnard. 
Because there were at the time no academic institutions with large- 
scale digital zine projects, I reviewed three sites run by private groups. 
I also reviewed two academic digital collections that feature radical or 
obscure publications. 

My analysis revealed several qualities that most online exhibits of 
zine-like material share. 

Asset management: (l) All five sites included full-color image 
scans with legible text and graphics; (2) All but one site included an 
option to download the asset in PDF form; (3) Four out of five sites 
included descriptive metadata about subject, author, and publication 
date to aid in discovery and to give context to the asset 

Navigation: (1) Every site evaluated had a descriptive homepage 
and a consistent look and feel; (2) All five sites utilized global 
navigation on each page to keep the user oriented. 



DIY Digitization 11 



Search and discovery: (l) Four out of five sites allowed users to 
browse for a zine by title; (2) Four out of five sites offered a keyword 
search function; (3) None of the sites offered a search by author or 
issue number function; (4) Four out of five sites made searching or 
browsing for a known-item simple and pleasurable. 

Tools and customization: Every site evaluated offered a "printer- 
friendly" version of their assets 

Aesthetics and usability: (1) Every site took care to ensure that all 
links and functions worked as they were expected to; (2) Four out of 
five sites used some type of backend content-management system to 
organize assets; (3) For the qualities adopted by all sites evaluated, I 
attempted to include them. 



Site Name and URL 


Launch Date 


Assets 


Zine Library.net 
httD://www.zinelibrarv.info/ 


None given 


"hundreds" of 
zines 


The Queer Zine Archive Project 
http://www.azap.orq 


Nov 2003 


154 issues 


Punk Zine Archive 

http://www.operationphoenixrecords.com/ 
archivespaqe.html 


2004 


120 issues 


Ling Long Woman's Magazine @ 
Columbia University 
http://www.columbia.edu/cu/lweb/diaital/ 
collections/linqlonq/index.html 


2005 


241 issues 


Anarchism Pamphlets in the Labadie 
Collection @ The University of Michigan 
http://www.lib.umich.edu/spec- 
coll/labadie/ 


1999 


478 pamphlets 



Figure ZINE-3: Sites Evaluated for Competitive Landscape Analysis 



There were other qualities present in some sites but not in others. 
These included RSS feeds, customizable user accounts and high-tech 
page turners. Because these qualities appeared in only some sites, I 
considered them to be optional for my site. 



12 DIY Digitization 



Interestingly, none of the sites offered any curriculum or 
supporting finding aids that would add necessary context to the 
materials. I planned to include lesson plans and a bibliography to 
accompany my zine scans. 

Creating a sitemap and wireframes. 

I first sketched wireframes for my site using paper and pencil, 
then translated those sketches into digital files. The wireframes turned 
out to be ambitious, and due to time and skill constraints, I was forced 
to scale down my original vision, but the creation of the sitemap and 
wireframes helped me synthesize all my ideas for the site into one 
visual presentation. 




Figure ZINE-4. Final site map for the exhibit. 



Generating Content 

Scanning and digitization. 

Ideally, for this project, preservation-quality scans should be 
made of each zine, along with its entire contents at the highest 



DIY Digitization 13 



possible resolution in addition to any presentation and/or thumbnail 
versions. However, I was limited in my ability to create high- 
resolution scans due to several factors. The first issue was the 
constraints of the hardware and software at my disposal. I worked on 
scanners and computers that Barnard undergraduates have access to 
in the library's computer lab. More sophisticated equipment might 
have resulted in better scans. 

The condition of the zines themselves also contributed to poor 
resolution. Because most zines are produced using cheap materials, 
many of the oldest were beginning to deteriorate as the paper 
yellowed and the ink faded. Additionally, poor photocopying 
resolution in the original zine made some of the digitized page images 
appear grainy and pixilated. 

After much experimentation, I balanced preservation and 
presentation needs with time and resource scarcity by scanning each 
zine once at 600 DPI or higher. Next, I saved two JPG versions of each 
zine: a presentation copy at a maximum height of 6oopx and a 
thumbnail copy at a maximum height of 90px. I preserved the aspect 
ratio of each scan each time I resized the image. When necessary, I 
used image editing software — either the open-source GIMP 
(http://www.gimp.org) or Adobe Photoshop, depending on what was 
installed on the computer I was working on that day - to tweak the 
contrast levels of the scan and improve legibility. For each zine's 
cover, I created a slightly larger thumbnail which is presented on the 
Home page and the "Featured Zines" page. The last step was to 
convert all of the JPGs to PDF and create a print version of each zine 
for users to download. 

These digitization decisions allowed me to produce legible copies 
of each zine while maximizing disk space. Because I didn't have access 
to an image database or a content management system, I simply 
organized all of the files in a series of folders on my desktop, giving 
each file a descriptive name following a clear convention. When the 
exhibit was complete, these folders were uploaded, along with the 
HTML and CSS files to the Barnard server. 



14 DIY Digitization 



Generating metadata. 

I was able to take the metadata for the exhibit from the existing 
OPAC records. Each zine is currently assigned cataloged in a MARC 
record as part of the Barnard/Columbia joint OPAC, CLIO. Current 
metadata includes: title, an author or creator (when applicable), 
physical description, a publisher and date, Library of Congress subject 
headings and a summary or abstract. For each zine's gallery page, I 
used only the author, title, summary and call number fields. 
Additionally, I included a link to CLIO so users could locate the zine, 
check on its availability, and order it through interlibrary loan. I was 
able to add additional metadata about individual zinesters who 
responded to my copyright requests, including links to each zinester's 
current projects or personal websites. 

Writing Original Content 

A major component of the site was the contextual material that 
would add value and meaning to the zine scans for users. To meet this 
need, I wrote an "About" page describing the exhibit, as well as a "For 
Teachers" page that included: 

• a brief explanation of why zines make good teaching tools 

• three essential questions related to zines which could be used 
to frame curriculum planning 

• A list of suggested resources for educators 

The most time consuming content pieces to develop were the 
three lesson plans designed to help educators teach with the zines in 
the exhibit. Drawing on my background as a teacher, I designed these 
lesson plans around essential questions related to media literacy and 
social science content areas, then aligned them to three different 
learning levels: intermediate, secondary, and post-secondary. The 
lesson plans are student-centered and challenge students to interact 
with the zines in the exhibit through discussion and evaluation. 

Building and Testing the Site 

Ideally, the user interface for any site should be intuitive and 
promote discovery. For this project, I was limited to very basic web 



DIY Digitization 15 



design software and programming languages that have a low-barrier 
to mastery, but I was determined to make the site as usable as possible 
given the constraints. 

To build the site, I again used hardware and software in the 
Barnard Library computer lab. I first attempted to build the site using 
Microsoft FrontPage, and then switched to an open-source HTML 
editor called Mozilla Kompozer (http://www.kompozer.net). Both 
FrontPage and Kompozer have "WYSIWYG" interfaces - an acronym 
for "what you see is what you get" - in that they allow users to create 
web pages using an interface that mirrors how the final product will 
appear (Myers, 1998). I needed to have a firm grasp on what was 
happening in the actual code behind my pages as I made changes. 
Additionally, I wanted to be able to customize my site, and the 
templates available in FrontPage and Kompozer felt limiting. 

After a few days of struggle, I settled on developing the code for 
my site by hand using the simple text editing software available on 
most computer operating systems. The flexibility of being able to work 
on my files from any location made this project much easier to 
complete in only 10 weeks. I created and edited my files in Microsoft 
Notepad at Barnard Library, and could easily open them at home on 
my Apple laptop using either TextEdit or a free trial version of the 
excellent application, Coda (http://www.panic.com/coda/) which 
combines a WYSIWYG interface with an easy-to-use text editor. 
During the initial coding phase, I began by building a framework for 
each site using a common layout, menu bar, and footer using simple 
HTML tags such as those for images, links and tables. I also selected a 
patriotic red-white-and-blue color scheme and created an image 
banner to run along the top of the page that included the site's title in 
a typewriter-style font and some randomly-placed stars to mimic a 
zine-like feel. Later, the color scheme was changed to a more punk- 
inspired pink and black, to better reflect the lack of overt patriotism 
expressed in the zines themselves. 

Once a basic page template was complete, I created a Cascading 
Style Sheet (CSS) file which governed the look and feel of each page 
and made coding the rest of the site easier. Using a CSS file is a simple 



16 DIY Digitization 



way to add style (such as fonts, colors and spacing) to every page in a 
website without having to edit each page individually (Bos, 2010). In 
order for the CSS file to work, I added a line of code in each page's 
HTML file that directed a user's web browser to "link" to my style 
sheet file, called "text.css". This file dictated the background and font 
colors of each part of each page, the margins for the different dividers 
and tables, and even the spacing of thumbnail images in my photo 
gallery. Before beginning this project, I had never worked in CSS. 
However, I found the language simple to learn and easy to use once I 
understood the fundamentals. As I worked, I referenced countless 
tutorials and open-source code available on the web to hack my way 
through the rest of the coding process. 

The biggest coding challenge was creating a workable photo 
display gallery to present my zine scans. During the site mapping and 
wire-framing (see Figure 8), I had determined that I wanted users to 
interact with thumbnail versions of a zine's pages and select which 
pages they'd like to see in a larger, presentation-quality view. At first, I 
tried copying an open-source photo gallery HTML file that I found 
online. This allowed me to build the bare bones of my page layout, and 
indeed included thumbnail images and presentation-quality views. 
However, the photo gallery had one weakness: every time I moved my 
mouse off of the thumbnail I wanted to view, the presentation-quality 
view disappeared! This would have made my site very difficult to use, 
as it was impossible to scroll, zoom, or even save the presentation- 
quality view while still keeping your mouse over the thumbnail view. 
Finally, a programmer friend-of-the-collection helped out by 
producing a small piece of JavaScript code that allowed me to keep the 
presentation-quality view open without having to keep my mouse 
perfectly still. This code allowed me to create a version of the site that 
I was excited to test with users. 

Usability Testing 

Before going live, I tested the site's usability with three different 
test subjects, each representing a different group of Barnard Library 
stakeholders. Each test subject was given the same set of tasks to 



DIY Digitization 17 



complete. Figure ZINE-5 lists the tasks administered and whether or 
not the subject was able to complete the task without guidance. 



o a 

o 3 

3 0 

CL 



<< 
-I ■■ 

Cfl 



c 


O 




'\'1 






Q 


? 




O 




w 




a 

ZL 


izr 

■~ 


— - 

O 




=r 

w 


0 




3 












JJ 








=T 




■ ^ 








■1) 


^ 




< 




J" 




3 
~ 1 


□ 


•ET 


!= 






■1/1 


3 




a 








0 


a. 


■< 








Q. 






5 






=1 
















0 




j? 


0 
3 






















_— 1 






3C 

























=: 

O =± 



Hi 

T3 
ITS 

o 

-*-• 
ID 

Z: 

1— u 

5 — —1 

er O 

A> <*> 



:= 
■< 

N 
Zt' 

:= 

cd 



Do 




r— 

c 


IJAV 




0 
cu 

ft 








Z> 


ii 




ft 












3 


0 




Si 




c 


0 






z\ 


CD 


&i 


"C 




C" 


1; 


c 


<C 


3-' 


cz 


CE 


CE 




U 
c 


6 


how 



3 

CO 



R 



CD 



15 
1= 
—•■ 

n 

c 
c 



c 

CD 



CE- 



CE' 



ZT 

Cfi 

CE 
H 
■ 

o' 
r 

CI 
CE 



CP 



m a Q 
— 3. 



CD 



5' 



o 

5t' 
=r 

C 

cr 

3 
_j 
to 

O 

— 

3' 
=> 
m 



o 
cz 

o 

=1 =; s 



CD 
PI 



o 

c 



CE' 
CD 



n 



tE' 

CD 



O CD 



CP o 
M C= 

=■ to 



a 



is: 



cz 



ft' 
Cn 

CD' 



Figure ZINE-5: "Exhibits and Protest" Site Usability Tasks and Results. 



18 DIY Digitization 



In addition to identifying tasks that would be difficult for users to 
complete, I also made anecdotal records of the test subjects' 
comments and feedback. Based on the results of the testing, I made 
the following improvements to the site: 

Increased the size of all fonts used by lpx 

• Added a mouse-over function to each zine cover image on the 
homepage that listed its title to aid in identification 

• Added a "download all" link to a PDF containing all image files 
associated with each zine 

Made the copyright documentation on each zine more 
prominent 

Launching and Publicizing 

Upon completion of testing, the final version of the online exhibit 
(http : / / www.barnard.edu /library/ zines / exhibits / online / 
elections/index.html) was launched on July 23rd, 2008. A link to the 
exhibit was posted on the Barnard College Library homepage, and a 
blog post about it was added to the institution's Livejournal 
(http://barnardzines.livejournal.com/). Emails were sent to a zine 
librarians' listserve and to other contacts and friends of the Zine 
Collection. I also sent a link to the exhibit to colleagues and friends in 
K-12 schools across the country. Since the initial launch, Freedman 
has continued to publicize the online exhibit in her talks and outreach 
activities for the collection. 

Results and Next Steps 

Informal evaluation of the project demonstrates that it has begun to 
meet its goals. User feedback on the site has been overwhelmingly 
positive. By making Barnard's zines accessible on the web to millions 
of people across the globe, the exhibit has indeed improved access to 
the collection. In an age when discovery on the web is primarily done 
through Google or other search (Belden, 2008), it is encouraging that 
search terms such as "zine lesson plans," "zines and elections," and 
"teaching with zines" consistently return the site in the first page of 



DIY Digitization 19 



search results. This is an indication of how many sites continue to link 
to the exhibit since its launch. 

It is obvious from the Access and Use Survey that such an effort 
is both desired and respected by stakeholders. The exhibit only 
scanned selections of the zines featured, not entire issues, it may not 
be seen to contribute to the long-term preservation of the individual 
artifacts. However, creating a home on the web for zines around a 
contemporary issue can be seen to be contributing to the long-term 
preservation of the genre by making zines relevant in the digital age. 

It remains to be seen whether this site will indeed drive users to 
the Zine Collection's other resources. Further evaluation should be 
done in order to determine whether or not this exhibit is directly 
contributing to increased access or use of the collection by 
Barnard/Columbia community members, outside researchers, K-12 
educators, and other stakeholders. Repeating the Access and Use 
Survey annually may be a step in that direction. 

Long-term sustainability and continued effectiveness of the 
exhibit are an issue. Freedman or future interns will need to take on 
the responsibility for maintaining and updating the exhibit as 
necessary over time. As Barnard Library further develops its web 
presence, the exhibit's look, feel, and even its content could become 
outdated. At this time, however, there is no reason why the exhibit 
cannot stay live for the foreseeable future without financial cost or 
significant staff time commitment. In order to maximize the exhibit's 
effectiveness, Barnard could consider: 

Continuing to promote and publicize the current exhibit, 
focusing on alternative outlets such as Wikipedia, educator 
websites, and media literacy blogs 

Developing an evaluation plan to determine the impact of the 
online exhibit on the stated project goals. 

• Securing site analytics data on page usage and download stats 
to measure usage of the exhibit, as well as effectiveness of 
marketing techniques 



20 DIY Digitization 



Due to the support for this project from users, Barnard could 
consider digitizing more of the collection. Next steps could include: 

Creating more online exhibits around themes or subjects of 
interest to stakeholders if this first exhibit proves valuable 

Creating more subject guides, lesson plans and bibliographies 
about zines and zine history, and making them available online 

Collaborating with Columbia's New Media Teaching and 
Learning group in order to ensure that the user interface 
promotes teaching and learning with zines as primary sources, 
art objects and media literacy teach tools. 

An open question is whether or not Barnard should move forward 
with digitizing the entire collection. Although this project was able to 
overcome many barriers to digitization of the genre (Wooten, 2009), a 
larger-scale project might open the door to more difficulty with 
copyright, permissions, privacy, and preserving the user experience of 
interacting with a zine's physical form. 

Conclusion 

This project showed that it is possible to create innovative web 
resources for a variety of stakeholders with a minimum level of 
technological and know-how. It provided a great opportunity for 
Barnard to continue to lead in the field of zine librarianship. This 
online exhibit supports teaching, learning, and research with quality 
and findable digital assets that highlight Barnard Library's strengths. 
With the current low barrier to web authorship, it is not enough for 
academic special collections to simply have a web page. Instead, more 
libraries and institutions can take advantages of the resources 
available to them - whether it is an eager intern, an exciting 
collection, or a timely theme -to create a resource that will meet the 
needs of users and make their holdings accessible to and available for 
generations to come. 



DIY Digitization 21 



References 

Astbury, K. (2006). French theatre of the first empire: Enhancing 
research-based learning. Warwick Interactions Journal 28. 
Retrieved from http://www2.warwick.ac.uk/services/cap/ 
resources / pubs/interactions / archive/issue28 / abastbury/ astbury 

Belden, D. (2008). Harnessing social networks to connect with 
audiences: If you build it, will they come 2.0? Internet Reference 
Services Quarterly, 13(1), 99-111. 

Bos, B. (2010). Cascading style sheets. W3C Consortium. Retrieved 
from: http://www.w3.0rg/Style/CSS/ 

Congdon, K. G., & Blandy, D. (2003). Zinesters in the classroom: 
Using zines to teach about postmodernism and the communication 
of ideas. Art Education. 56(3), 44-52. 

Daly, B. O. (2005). Taking whiteness personally: Learning to teach 
testimonial reading and writing in the college literature classroom. 
Pedagogy, 5(2), 213-246. 

Dempsey, L. (2006). Libraries and the long tail: Some thoughts about 
libraries in a network age. D-Lib Magazine, 12(4). 
doi:io.i045/april20o6-dempsey 

Dodge, C. (2008). Collecting the wretched refuse: Lifting a lamp to 
zines, military newspapers and Wisconsonalia. Library Trends, 
56(3), 667-677- 

Duncombe, S. (1997) Notes from underground: Zines and the politics 
of alternative culture. New York : Verso. 

Jones, M. (2008). Elections and protest: Zines from the Barnard 
Library collection. Barnard Library Website. Retrieved from 
http : / / www.barnard. edu /library/ zines / exhibits/ online/ elections / i 
ndex.html 

Lesk, M. (2007). From data to wisdom: Humanities research and 
online content. Academic Commons. Retrieved from http:// 
www.academiccommons.org/ commons/essay/ michael-lesk 

Liu, S. (2008). Engaging users: The future of academic library web 
sites. College & Research Libraries, 69(1), 6-27 



22 DIY Digitization 

Myers, B. (1998). A brief history of human computer interaction 
technology. ACM interactions, 5(2), 44-54. Retrieved from: 
http://www.cs.cmu.edu/~amulet/papers/uihistory.tr.HTML 

Soderberg, J. (2002). Copyleft vs. copyright: A Marxist critique. First 
Monday, 7(3). Retrieved from: http://ojphi.org/htbin/ 
cgiwrap/bin/ojs/index.php/fm/article/viewArticle/938/860 

Wan, A. J. (1999). Not just for kids anymore: Using zines in the 
classroom. Radical Teacher, 55, 15-19. 

Wooten, K. (2009). Why we're not digitizing zines. Duke University 
Libraries Digital Connections Blog. Retrieved from 
http:/ /library.duke.edu/blogs/digital- 
collections /2009/09/21/ why-were-not-digitizing-zines / 

Wright, F. (1997) The history and characteristics of zines. TheZine & 
E-Zine Resource Guide. Retrieved from 
http : / / www. zinebook. com/resource/ wright 1 . HTML 



Digitizing Civil Rights: An Omeka-based 
Pilot Digital Presence for the Queens 
College Civil Rights Archive 



Valery Chen, Jing Si Feng, Kevin Schlottmann 
(Queens College, CUNY) 



Abstract 

The Queens College Civil Rights Archive of the Department of Special 
Collections partnered with the Queens College Graduate School of 
Library and Information Studies to create a pilot web presentation 
using the open-source Omeka platform. Phase I of the project, 
conducted during the Spring 2010 semester, outlined a method for 
institutions of limited means to enter into the world of digitization 
using existing resources while highlighting the difficulties involved 
with metadata and IT support, and the advantages of involving 
graduate students. 

Keywords: Civil rights, Digitization project, Omeka, Plug-ins. 



Introduction 

In the spring of 2010, the Queens College Civil Rights Archive of the 
Department of Special Collections partnered with the Queens College 
Graduate School of Library and Information Studies to create a pilot 
web presentation using the open-source Omeka platform. The 
synergy created between the technological skills found in the library 
school faculty and student body and the desire of the Civil Rights 
Archive to begin digitization of key holdings allowed the rapid 



23 



24 Digitizing Civil Rights 



creation of a powerful web presentation platform. The process also 
outlined a method for institutions of limited means to enter into the 
world of digitization. 

Queens College Civil Rights Archive 

The Civil Rights Archive of the Queens College Department of Special 
Collections and Archives collects published and unpublished works 
relating to civil rights activities such as personal papers, community 
materials, organizational records, non-print materials, and artifacts. It 
also conducts oral histories to supplement its collections. The archive 
is particularly strong in materials documenting civil rights work by 
Queens College students during the early 1960s. The Archive seeks to 
provide evidences of the under-documented Northern involvement in 
the civil rights movement. 

The Archive was founded in late 2008 around an estimable 
collection of personal papers donated by alumnus Mark Levy. Since 
then almost a dozen other personal collections relating to civil rights 
work in the 1960s have been donated by College alumni, and the 
Archive continues to actively collect in this area. 

Queens College Graduate School of Library and 
Information Studies 

The Queens College Graduate School of Library and Information 
Studies prepares library/information service professionals to meet the 
information and literacy needs of the New York metropolitan region 
and beyond. It is the only American Library Association accredited 
program for library and information studies within the City University 
of New York. The school prepares graduates to serve a broad segment 
of the metropolitan area's multicultural, multiethnic and multi-lingual 
population in a variety of institutional and informational settings. 
Through research, publication and other forms of scholarly activity, 
the school contributes and transmits new knowledge to society and 
the profession. The faculty provides opportunities for students to 
attain the competencies needed to participate in the evolving 



Digitizing Civil Rights 25 



electronic age by providing a technologically rich teaching/learning 
environment. 

Project Origin 

The Queens College Department of Special Collections, wherein the 
Civil Rights Archive is located, was acutely aware of the need for its 
collections to have a digital presence. It will soon be true that archival 
materials that are not electronically accessible in some way, whether 
via an OPAC or on website, will be no better served than in a dark 
archive. Given the limited resources of an urban public university, the 
Department had been unable to secure sufficient financial and 
technological support for an independent digitization project. Head of 
Special Collections Dr. Ben Alexander is also teaching in the Graduate 
School of Library and Information Studies, and he approached Dr. 
Kwong Bor Ng to discuss a mutually beneficial way to begin the 
process of building a digital presence. Drs. Ng and Alexander decided 
to expand the Special Collections Fellowship program, which provides 
archival graduate students at Queens College with a broad range of 
professional archival experience, to include a technology component. 
Dr. Ng selected two graduate students with extensive coding skills to 
do the actual work of creating an Omeka presentation website. The 
hope was to create a mutually beneficial arrangement: under Dr. Ng's 
supervision, the two graduate students were able to gain real-world 
experience in building an Omeka platform, while the Civil Rights 
Archive was able to lay sufficient groundwork to seek grant funding in 
support of a larger digitization project. 

Staffing/Workflow 

Drs. Alexander and Ng served as project coordinators. The semantic 
team, which was co-extensive with the Department of Special 
Collections staff, consisted of Dr. Alexander, Archives Adjunct Katie 
Hughes and Archives Assistant Kevin Schlottmann. The technical 
team was headed up by Dr. Ng, who supervised two of his technology 
graduate students, Valery Chen and Jing Si Feng. They were given 
academic credit as independent study students and Special Collections 



26 Digitizing Civil Rights 

Fellows to build the Omeka website. The semantic team was 
responsible for selection, digitization, and metadata creation. The 
technological team was responsible for the installation and 
development of the Omeka presentation. Both teams were involved in 
the development of the Dublin Core metadata schema, and they also 
collaborated in creating the user experience of the website. 

The project began with a meeting in February 2010. This meeting 
was initially a brainstorming session, but the teams were able to agree 
to the basic semester goal of a pilot website, as well as a rough 
timeline. Once the semantic team completed selection and the 
metadata schema was ready, a few items from the Civil Rights Archive 
were digitized per week and forwarded to the technological team. 
Meanwhile, the technological team was preparing the Omeka website 
for import of digital items. This continued as an iterative process for 
the entire Spring 2010 semester. At biweekly meetings, the website 
and the digital items were discussed, and both were constantly 
improved. 

Implementation - Semantic Team 

The semantic team, consisting of the Department of Special 
Collections staff (Department Head Dr. Benjamin Alexander, Archives 
Adjunct Katie Hughes and Archives Assistant Kevin Schlottmann) was 
responsible for selection, digitization, and metadata creation. 

Selection 

When the semantic team began discussing what to digitize, it 
considered materials from the Civil Rights Archive, the College 
Archive, the Performing Arts collection, and the Rare Book, Zine, and 
Artists Book collections. It quickly became clear that the civil rights 
materials were best suited for this pilot project, for reasons such as 
processing status, donor relations, fitting into the Archive's specific 
mission of engaging with the broader community, copyright status of 
the materials, and the attraction of having students continue to work 
with the material. 



Digitizing Civil Rights 27 



First and foremost, the Civil Rights Archive had the best- 
cataloged collections in the Department. The majority of the Archive's 
holdings are fully processed, have finding aids, and are under archival 
control. This level of processing also allowed for existing contextual 
information to be used by the teams. 

The civil rights materials are among the most prized and high- 
profile holdings in Special Collections. The recent founding of the 
Archive garnered attention from campus and activist communities, 
and received local press coverage as well. From a donor relations 
viewpoint, digitizing these collections would generate goodwill within 
the alumni activist community that contains future donors and 
supporters. The Archive's living donors are still very interested in 
their materials and legacies, and have repeatedly expressed strong 
interest having their materials digitized. 

The semantic team also conducted a review of existing civil rights 
digital archives, which showed many worthy efforts already underway. 
The University of Southern Mississippi, which holds one of the largest 
archival collections about civil rights work in Mississippi, created the 
Civil Rights in Mississippi Digital Archive, an "Internet-accessible, 
fully searchable database of digitized versions of rare and unique 
library and archival resources on race relations in Mississippi" 
(University of Southern Mississippi Special Collections, 2006a). The 
Civil Rights Digital Library, hosted by the University of Georgia, is "a 
partnership among librarians, technologists, archivists, educators, 
scholars, academic publishers, and public broadcasters" that provides 
federated searching of digital civil rights materials from almost 100 
different institutions (Digital Library of Georgia, 2009). 

These two excellent examples, among dozens of others, illustrated 
two major reasons why the semantic team chose to digitize materials 
from the Civil Rights Archive, rather than from other special 
collections. First, it was found that there exists a vigorous online 
community that any institution holding archival civil rights materials 
must join to remain relevant and accessible. Second, the team found a 
paucity of material that relates specifically to Northern contributions 
to the civil rights movement, and thus digitization of the Queens 



28 Digitizing Civil Rights 



College Civil Rights Archive would add new perspectives to the online 
community, in keeping with the Archive's stated mission of engaging 
the broader archival civil rights community. 

The copyright status of the many of the materials in the Civil 
Rights Archive was clear, because photographs and personal papers of 
known provenance could be easily cleared by the creators with whom 
the Department has a relationship. Many of the other areas in special 
collections have a murkier copyright status, which is a major potential 
impediment to digitization. 

Finally, the civil rights materials were primarily processed by 
archival students from the Queens College Graduate School of Library 
and Information Studies as part of the Special Collections Fellowship 
program. It seemed natural to continue the collaborative effort 
between the library school and Special Collections by having the 
student-processed materials brought into the digital realm by Fellows 
as well. 

Digitization 

The actual scanning of the items was not the focus of this pilot 
project. A proper scanning procedure that will create archival images 
in the TIFF format, such as that developed by the University of 
Southern Mississippi (University of Southern Mississippi Special 
Collections, 2006b), will be developed during Phase II in conjunction 
with implementation of a digital asset management system. The items 
digitized for this project were scanned with an Epson 10000XL 
scanner using Adobe Photoshop CS3. After cropping, deskewing, and 
adjustments to contrast and level, the images were saved as 300-dpi 
Web-optimized jpegs and provided to the technical team. 

Metadata 

The development of a robust metadata schema was a primary goal 
of this project. The semantic team examined a range of available 
schemas, such as METS, MODS, and PREMIS, but it very quickly 
became clear that Dublin Core (DC) was ideal for a variety of reasons. 
Unqualified DC is compatible with basic Omeka; it is a simple and 
easily-understood schema; it can be extended by using qualified DC; it 



Digitizing Civil Rights 29 



is well-established; and similar projects are using it. The latter was a 
particular influence, both because the Civil Rights Archive hopes to 
digitally collaborate with other institutions as well as because there 
are a wealth of relevant resources available. The two key sources used 
by the semantic team were published DC schemas from the University 
of Southern Mississippi and the North Carolina Exploring Cultural 
Heritage Online project (Graham & Ross, 2003; NC ECHO, 2007). 

The semantic team created a detailed qualified Dublin Core 
schema, but after much discussion the teams decided to work with an 
unqualified schema, because that was the Omeka default. The 
semantic team continued to create qualified DC, so that in the future 
the project will be able to implement this more detailed schema. 

The project utilized many controlled vocabularies. Library of 
Congress Name and Subject Authorities were used for person and 
subject terms; DCMI and IAMA for type and digital format, 
respectively; and ISO 8601 for dates. For the analog Medium field, 
the Getty's Art and Architecture Thesaurus proved most useful. 
Geographic data were placed in the Coverage field. LC subjects were 
used for the general geographic area, such as the town or state, while 
latitude and longitude data were taken from Google Maps by manually 
entering a known address and harvesting the geospatial data 
provided. The technology team was able to use the latitude and 
longitude data to create a Google Map reflecting the geographic 
location of the digitized items. 

The teams both felt it important to offer maximum searchability, 
and the semantic team thus also provided the full text of digitized 
items, using OCR software. This proved to be a time-consuming 
additional step, in particular the proofreading of the computer- 
generated text. The print quality of some materials was quite poor, 
and the many drawings and photographs were also difficult for the 
software to interpret. Better software and more experience creating 
OCR text should make this easier as the project moves forward. 

Omeka also has the ability for users to add tags. The teams 
decided to take selected controlled-subject values from the Dublin 
Core metadata and use them as tags as well, to allow testing of 



30 Digitizing Civil Rights 



features such as the tag cloud. The teams engaged in an interesting 
discussion about how tags would be used in this project - some 
wanted to keep the vocabulary controlled, while others wanted to 
encourage users to add tags as they saw fit. The teams decided to 
allow tags to be used as user-generated metadata, in keeping with the 
Web 2.0 spirit of Omeka. 

The metadata was created manually by the semantic team in an 
MS Word table, and transferred into Omeka by the technology team. 
A future goal is an automated process for metadata transfer. 

Implementation - Technology Team 

The technology team for the project consisted of two independent 
study graduate students, Valery Chen and Jing Si Feng, as part of the 
Special Collections Fellows program, and their instructor and project 
coordinator Dr. Ng. These Fellows began working on the project after 
the first meeting with the semantic team on February 16, 2010. Over 
the next ten weeks, the Fellows downloaded, installed, and modified 
Omeka, an open-source web-publishing system, and completed the 
first phase of the project on May 13, 2010. 

Why Omeka? 

The purpose of this project was to create a web presentation to 
showcase the unique and valuable holdings of the Queens College Civil 
Rights Archive, and at the same time provide the Fellows an 
opportunity to learn how to build a digital archive using a web- 
publishing system. In any project, it is important to consider the use 
of proprietary system versus nonproprietary/open-source. Omeka is 
relatively a new software package that describes itself as a web- 
publishing platform on its website: 

Omeka is a free, flexible, and open source web-publishing 
platform for the display of library, museum, archives, and scholarly 
collections and exhibitions. Its "five-minute setup" makes launching 
an online exhibition as easy as launching a blog. Omeka is designed 
with non-IT specialists in mind, allowing users to focus on content 
and interpretation rather than programming. It brings Web 2.0 



Digitizing Civil Rights 31 



technologies and approaches to academic and cultural websites to 
foster user interaction and participation. It makes top-shelf design 
easy with a simple and flexible templating system. Its robust open- 
source developer and user communities underwrite Omeka's stability 
and sustainability. (Omeka, 2010, Project section.) 

Omeka is an open-source web-publishing system developed by 
the Center for History and New Media at George Mason University. 
According to the Omeka web site (Center for History and New Media, 
George Mason University, 2010) Omeka is easy to install, allows great 
flexibility for customized web interface, and supports multiple 
plugins. All these features were appealing for this project. 

Another compelling reason to choose Omeka was the potential 
inherent in the exhibit feature. Omeka has the Web 2.0 ability of 
allowing users to create their own exhibits from the digital collections. 
A primary goal of the Civil Rights Archive is to engage the educational 
community and encourage use of its materials. By providing digital 
surrogates and contextual information, this website would allow a 
teacher or professor to tailor their use of the materials in an exhibit, 
and also make them accessible to other educators seeking similar 
uses. This type of educational contextualization was a key reason to 
digitize the collection. 

Description of Phase I (Feb 16, 2010 - May 13, 2010) 

This section discusses the installation of Omeka, the addition of 
various plugins, the details of the most heavily manipulated pages, 
and examines particular technical problems and solutions 
encountered during Phase I of this project. 

Installation 

Initially, separate Omeka instances were created for each Fellow 
on the technology team to experiment independently. After both 
instances were adequately developed, the best features were selected 
from each and transferred over to a new Omeka installation, running 
on version 1.2. 

Omeka l.i, the latest version available at the time of the first 
installation process, was downloaded by the technology team and 



32 Digitizing Civil Rights 



installed. The Omeka system was in a zip file, and the technology 
team had to unzip the file to extract all the necessary files for 
installation. Each Fellow downloaded the zip file and unzipped the 
files successfully. The next step was to connect to the server remotely 
and upload the files for installation. Omeka consists of thousands of 
files, and Adobe Dreamweaver could not handle such a massive 
upload. Using an FTP client, such as FireFTP, was found to be the 
best practice. FireFTP supports large uploads and does not terminate 
in the middle of an upload. If termination does occur, FireFTP 
automatically reconnects to the server to continue with the upload. 

The Omeka installation folder in the directory was removed by 
bash shell script for security. Administrator and Super accounts were 
created, and the system was up and running. 

Creating Items, Tags, and Collections 

A collection can be created in the Omeka Admin page by filling in 
the name and description of a collection. An item can be created by 
filling in the Dublin Core fields in the Omeka Admin page and adding 
the item to a collection. Tags can be added to each item to create 
more access points. 

All image values in the Omeka General Setting should be defined 
before importing any image files. The values Fullsize Image Size, 
Thumbnail Size, and Square Thumbnail Size are crucial for Omeka to 
generate image output. Omeka automatically generates full size image 
and thumbnails during item creation. The technology team decided to 
change the full size image output and thumbnail size in the middle of 
the project, thus resulting in two different image sizes throughout the 
site. This meant that all the files needed to be uploaded again at the 
conclusion of the project, to ensure uniform image sizes. 

Selecting a Theme 

The public interface of Omeka was controlled by the files inside 
the "themes" folder. The Super can log on to the Omeka Admin page 
to choose a desired theme. More themes can be downloaded from the 
Omeka website and uploaded to the server. Both Fellows selected a 
different theme for their individual pilot Omeka sites. One chose 
"santa-fe" while the other chose "spring." 



Digitizing Civil Rights 33 



Plugins 

Many plugins are available for download from the Omeka 
website. The technology team installed Geolocation, Simple Pages, 
Dublin Core Extended, Dropbox, ExhibitBuilder, and Lightbox for the 
pilot site. Several outside interactive effects were also installed, 
including animated collapsible panels, text truncation, and a 
slideshow. Most of the plugins were easy to use and install without 
any hassles; the plugins specifically named above are discussed in 
more detail below. One plugin issue was that some of the plugins 
were written in plain JavaScript while others were written using 
jQuery. This often created clashes in the code as the dollar sign 
symbol ($) was used for different purposes in both JavaScript and 
jQuery. In JavaScript, the $ indicates a variable, while in jQuery the 
$ represents the start of a command. Since jQuery is technically a 
JavaScript library, the double meaning of the $ was problematic. 

Geolocation 

Geolocation uses features from Google Maps and allows users to 
view items in various geographic locations by clicking on the balloon 
pointers. The technology team decided to install this plugin because it 
gives users a visual, spatial representation of the materials in the 
archive. This visual aid can be useful for quickly identifying locations 
of particular interest, and it allows users an additional non-textual 
method of accessing digitized items. 

In order to use the Google Map functionality, the technology team 
first had to obtain a Google Maps API (application programming 
interface) key by creating an account with Google. Once the key was 
obtained, the team downloaded the packaged Google Map plugin from 
Omeka and uploaded it to the server. The key was then supplied 
through the administrator interface to activate the plugin. From the 
administrator interface, it is possible to customize the Google Map 
API. For example, the administrator can set the default location, 
which is controlled by latitude and longitude, and the zoom level 
ranging from 1 to 20, where 1 displays a view of the Earth and 20 
displays a street level map. 



34 Digitizing Civil Rights 



The geographic location was added manually through the 
administrator interface. The latitude and longitude were entered for 
each digital item's metadata. The technology team hopes to create a 
plugin in the future that will automatically locate the spatial location 
information from the Dublin Core metadata and supply it accordingly 
so that as items are uploaded they will automatically appear on the 
map. 

All items with geographical coordinates are represented by the 
default red pinWhen users click on the marker, the corresponding 
item will pop up in a balloon with limited metadata identifying the 
item. Users can then click on the balloon, which brings them to the 
particular item's display page. The team plans to enhance the map by 
creating a color-coding scheme that will correspond to an item's 
categorization. For example, items relating to a protest could have a 
red marker, while Freedom School items could have a blue marker. 
This functionality will add additional visual meaning to the map. 

Simple Pages 

Simple Pages provides the ability to handle html codes input by 
the Super at the admin page. The team used the Simple Page plugin 
to create the "About Us" section of the website. 

Dublin Core Extended 

Omeka defaults to Unqualified Dublin Core. A Dublin Core 
Extended Plugin can be installed to create more descriptive fields. A 
primary consideration when installing Dublin Core Extended is 
whether all the additional fields are necessary for a given project. Too 
many fields slow down the data entry process and scrolling time, and 
since all the Omeka data entries are web-based, a wrong click by the 
mouse can result in loss of data. The teams hope to take advantage of 
qualified Dublin Core in Phase II. 

Dropbox 

Omeka uses a web-based data entry system; each field needs to be 
typed in separately. A Dropbox plugin can be installed to facilitate 
bulk file uploads, allowing multiple files to be uploaded at the same 
time into a Dropbox Folder. When adding a digital item to the 



Digitizing Civil Rights 35 



collection, the user can then select that item from the Dropbox. 
Dropbox also supports the creation of multiple items, and allows bulk 
creation of Tags, Collections, and Titles. Other Dublin Core fields still 
need to be entered separately. 

ExhibitBuilder 

ExhibitBuilder was a more complex installation, as the 
architecture of the exhibit pages had to first be determined. The 
structure of the exhibit was defined as a Main Title Page with 
descriptions, a Sub-Section with descriptions, and individual pages 
with items in the layout of choice. The interface of an exhibit can be 
changed by accessing the screen, ess file under the ExhibitBuilder 
directory. 

Lightbox 

In order to enhance the user experience, a Lightbox was installed 
for viewing full size images. The Lightbox feature creates an overlay 
around an image and enhances the clarity of the image in the center. 
The latest version of Lightbox was downloaded from 
http://www.huddletogether.com/projects/lightbox2/, and installation 
process was closely followed from the Omeka Documentation page 
(Ebellempire, 2009). 

Heavily Manipulated files 

This section describes two particular areas of the Omeka platform 
that were heavily edited by the technology team. 

Item Display Page 

The default display page for each item is a non-stylized, 
exhaustive list of metadata that requires extensive vertical scrolling. 
Moreover, some of the DC and Omeka metadata are duplicative. In 
order to create a more user-friendly display of information, multiple 
files were heavily manipulated and outside plugins were incorporated 
to make the display page more interactive. 

The technology team found it challenging to determine which files 
control the item display page, because there are many files with the 
same name in different directories. However, once the structure of 



36 Digitizing Civil Rights 



how Omeka arranges its files was ascertained, the team was able to 
manipulate the code and customize the display. 

The first step after locating the correct file was to understand the 
predefined functions created by Omeka. This was achieved with the 
help from the documentation on the Omeka website. This stage was 
time consuming, as the team used a trial-and-error approach to learn 
how to use the different available functions. Once it was determined 
which function handled the selection of individual metadata, the team 
was able to control where each piece of metadata would be displayed. 

The display page was designed so that the most important 
information was provided in a clear, easy-to-read format. To 
minimize the scrolling, certain technical metadata was hidden from 
the user. The technology and semantic teams jointly examined each 
DC and Omeka data field, and decided on a short of list of fields to 
display. These included the collection, subject, tag, and rights fields. 
Creator, date, location and citation were included under the item's 
thumbnail image. Hidden fields included type, source, publisher, 
upload date, and extent. These fields were merely hidden - users may 
click on the "show" link to reveal the hidden metadata if desired. The 
show-and-hide functionality was made possible by installing jQuery, a 
lightweight JavaScript library, and the appropriate toggle-display 
code. The teams agreed that while an average user would not need to 
see this information on first viewing a page, more sophisticated users 
might want the option. The semantic team in particular found in its 
review of similar web projects that many sites did not include 
technical metadata, to the disappointment of archivists and librarians 
seeking to learn from how other institutions handle metadata. 

If a particular item had an accompanying full text, it was made 
available to the user. However, since a typical full text would cover an 
extensive area of the display page, the technology team used a 
truncation function to display a snippet of the full text along with a 
link to show more. This was also accomplished through jQuery. 

While most of the display page reorganization was controlled by 
the php files, CSS files controlled the final style touches, such as 
spacing, font sizes, and background color. 



Digitizing Civil Rights 37 



Homepage 

The default homepage also underwent a major redesign. Instead 
of the default static thumbnails representing the featured item, the 
team added a dynamic slideshow with captions that rotate through the 
featured items. The initial plan was to implement the slideshow using 
Adobe Flash. However, that idea was discarded as inefficient because 
new .swf files would have to be created for each new imaged added 
and each image taken down. Therefore, the best solution with the 
least amount of maintenance was to find a method to automatically 
pull images for display. The team achieved this by installing a jQuery 
slideshow. These images are thus easily maintained by a few lines of 
code. 

The technology team also added a right sidebar that allowed users 
to browse by various categories, including tag, date, location, subject, 
and type. The sidebar was initially occupied by a list of recently 
accessed items, but in conversation with the semantic team it was 
decided that creating multiple access points would be more useful to 
the user for navigation purposes. Currently, the list of categories is 
manually created through simple HTML code. The technology team 
hopes to automate the generation of categories in this sidebar, to 
simplify site maintenance. 

Goals and Discussion 

Phase II Goals 

The teams have many goals for Phase II. The stabilization and 
scalability of the project are a top priority, and this includes 
development of a scanning process and a streamlined metadata 
creation process, as well as finding stable server space and 
implementing a digital asset management system behind the Omeka 
website. 

In terms of the Omeka, the technology team hopes to add more 
sophisticated features to the map, such as item categorization. On the 
metadata side, it will implement qualified Dublin Core, and may 



38 Digitizing Civil Rights 



attempt to install the OAI-PMH Harvester. Finally, a long-term goal 
is the creation of a plugin that would allow use of TEL 

Discussion 

This pilot project identified issues that might be similar in other 
institutions with limited resources. The three key lessons from Phase 
I of this project relate to technical support, metadata, and the 
involvement of graduate students. 

Technical Support 

Proper server space is absolutely vital for any digitization project. 
The extremely limited IT resources available at Queens College mean 
that this pilot project is limited in its ability to grow beyond Phase I. 
This also impacts the Department of Special Collection's ability to 
implement a digital asset management system. If continued efforts to 
obtain College server space are unsuccessful, the project will examine 
other options such as approaching another CUNY school for a 
partnership, or perhaps even renting commercial server space. 

Metadata 

The semantic team spent the majority of its time and effort on 
metadata, both in developing the schema and creating the actual 
records. 

Creating metadata from scratch is extremely time-consuming. 
For each individual image, the team created titles and descriptions, 
applied a variety of controlled vocabularies, harvested geographic 
data, and proofread OCR text. With the current human resources, this 
process would be impossible to scale up. Thus, in Phase II the teams 
will attempt to streamline the process by automating some aspects of 
metadata creation. For example, technical specifications could be 
automatically imported and drop-down menus created for oft- 
repeated controlled vocabulary terms. Once a robust and well- 
described process is in place, Special Collections Fellows working on 
civil rights materials could be recruited to enter much of the metadata, 
which would provide the administrative team time to engage in other 
tasks, such as selection, quality control, and project planning. 



Digitizing Civil Rights 39 



Another issue the teams discussed was the extent to which 
existing archival description is helpful. As archival description by 
definition is not item-level, existing finding aids were only of limited 
use this pilot project, primarily providing contextual information. In 
the future however, incorporation of EAD finding aids would certainly 
allow additional types of categorization and searching. 

Bulk Metadata Creation in Omeka 

In order to facilitate the slow process of metadata entry for each 
item in Omeka, the technology team is experimenting with CSV 
Import in Phase II of the project. CSV Import, a plugin that handles 
bulk metadata creation, allows each Dublin Core field to be imported 
via a comma-separated file. However, more technical support is 
needed to define the PHP-CLO path setting with the server 
administrator. 

Graduate Students 

The collaboration of the Department of Special Collections with 
the Graduate School of Library and Information Studies was beneficial 
to both parties. Special Collections was able to lay the groundwork for 
future digitization and also to create a prototype Omeka website to 
show potential donors and partners, while the graduate Fellows 
gained valuable skills in actual implementation of a sophisticated 
software package. It should be noted that such a collaboration 
requires continuous interaction between and dedication from both the 
library school faculty and Department of Special Collections, to 
maintain standards and momentum as different classes of Fellows 
participate in the semantic and technology sides of the project. 

References 

Center for History and New Media, George Mason University. (2010). 
Omeka: about. Retrieved from http://omeka.org/about/ 

Ebellempire. (2009). Adding Lightbox to Omeka. Retrieved from 
http:/ / omeka.org/ codex/Adding_LightBox_to_Omeka 

Digital Library of Georgia. (2009). Welcome to the Civil Rights Digital 
Library. Retrieved from http://crdl.usg.edu/7Welcome 



40 Digitizing Civil Rights 

Graham, S. R. & Ross, D. D. (2003). Metadata and authority control in 
the Civil Rights in Mississippi Digital Archive. Journal of Internet 
Cataloging 6(10), 33-42. 

NC ECHO. (2007). North Carolina Dublin Core implementation 
guidelines. Retrieved from 
http : / / www. ncecho .org/dig/ncdc2007. shtml 

University of Southern Mississippi Special Collections. (2006). About 
the Civil Rights in Mississippi Digital Archive. Retrieved from 
http://www.lib.usm.edu/~spcol/crda/about.htm 

University of Southern Mississippi Special Collections. (2006). 
Guidelines for digitization. Retrieved from 
http://www.lib.usm.edu/~spcol/crda/guidelines/index.html 



Digitization on a Dime: How a Small 
Library and a Big Team of Volunteers 
Digitized 15,000 Obituaries in Just Over 
a Year 

Elizabeth Goldman (Kingston Frontenac Public Library) 



Abstract 

In 2006 and 2007, Chelsea District Library, a small public library in 
Chelsea, Michigan, digitized a collection of 15,000 obituaries on a 
small budget by using a staff of nearly 50 volunteers and open source 
software. The author describes the research and planning that led up 
to the project; unique aspects of the staffing and technology for the 
project; and the resulting database, which contributed to the library 
being named "Best Small Library in America" for 2008 by Library 
Journal. The database continues to be updated, expanded, and 
improved, and the use of volunteers as the primary workforce has had 
long term rewards for the library. The chapter demonstrates the 
accessibility of digitization projects to libraries, even those without 
pre-existing expertise, large staffs, or big budgets. 

Keywords: Database, Digitization, Genealogy, Obituaries, Open 
source, Volunteers. 



Introduction and background 

Especially in small towns, the public library may serve multiple roles 
as library, museum, archives, and community center. In many cases, 
libraries accept donations of rare or unique historical material from 



41 



42 Digitization on a Dime 



families in the area with little thought for long-term maintenance. 
Technological advances in the last few decades have given libraries 
new options for preserving local history collections and making them 
more accessible through digitization. While even the smallest libraries 
have staff educated in reference and circulation procedures, however, 
relevant training in archival methods and technology is rare, making 
the prospect of a digitization project overwhelming. At the same time, 
small public libraries often have little or no funding for such projects. 

Chelsea District Library is a single-branch public library serving 
14,000 people in southeast Michigan, about 50 miles west of Detroit. 
The library was established in 1932 by the local Women's Club and, in 
1999, became a district library system serving both the town of 
Chelsea (population 5,000) and surrounding townships. The earliest 
settlements in the Chelsea area date back to the 1830s and many local 
families can trace their roots back to the town's founders, resulting in 
a rich history. From its earliest days, the library served as a repository 
for local historical and genealogical material, housing a local history 
room on the cramped third floor of the McKune House, its location 
from 1959-2000. In 2006, after extensive renovations and additions 
to the McKune House, the library moved from temporary quarters 
back to its historic home on Main Street in Chelsea, leading to 
renewed interest in the local history collection. 

A collection of about 50,000 index cards known as the Family 
History Index made up a major component of the local history 
material. A retired lawyer and amateur genealogist named Harold 
Jones started the collection as a hobby, clipping obituaries from the 
local Chelsea Standard newspaper and other sources and pasting them 
onto 4x6 index cards, along with cross references that allow women 
to be located by maiden name. Upon his death in 1987, Jones' family 
donated the collection to the library, where it received extensive use 
by local and visiting genealogists. Library staff and volunteers 
completed a project from 2000-2002 to clean up and update the 
collection, since then volunteers have continued adding new clippings. 

As part of a planning process leading up to an election to fund an 
expansion of the building and the staff, the library identified local 



Digitization on a Dime 43 



history as a priority for Chelsea area taxpayers and the Family History 
Index as a prime candidate for digitization work. At the time, the 
professional librarian staff consisted of the director, three department 
heads (adult services, youth services, technology services), and a part- 
time librarian, leaving few resources to focus on a project of this 
scope. In the spring of 2005, the library replaced the departing part- 
time librarian with a full-time librarian (the author), adding 
additional duties of managing the library's website and digitizing the 
Family History Index. 

This paper will describe how a librarian and a team of four dozen 
volunteers completed the digitization of 15,000 records from the 
Family History Index in just over a year, resulting in a highly usable 
database that helped Chelsea District Library earn its distinction of 
"Best Small Library in America" for 2008 from Library Journal and 
the Bill and Melinda Gates Foundation. The first section will describe 
preparation for the project, including research in archival standards 
and digitization techniques, as well as the recruitment and training of 
the volunteer workforce. Section two will describe the decision- 
making process that went into the choice of open source software in 
order to create a user-friendly, free database of the records on a 
limited budget, as well as the work of developing and testing the 
database itself. Section three will offer an overview of the workflow for 
staff and volunteers as they did data entry, scanning, and 
proofreading of the records. Finally, the paper will describe the 
resulting database, current upkeep and expansion, and how the 
project served as a model for additional digitization work. 

Project preparation 

Research 

Initial research focused on archival and digitization standards. No 
one on the staff had a specific background in either archives or 
genealogy that would fit the requirements of the project, so research 
started from a very basic level. Research was conducted largely online 
and in books, as well as by speaking with archivists and libraries that 
had done similar projects. Internal research played a role, too: 



44 Digitization on a Dime 



understanding how the Family History Index had historically been 
used and its importance to the community; exploring the scope, size, 
and fragility of the collection; and agreeing on project goals. 

At the time, in mid-2005, quite a few libraries had launched 
efforts to put obituary indexes online for use by genealogists and 
historical researchers. The vast majority of these projects resulted in 
static webpages or simple databases that provided access to citations 
but not complete text. This early wave of digitization projects 
represented an important first step on the path toward full electronic 
access by giving researchers more complete information about the 
contents of a library's collection. However, those who found an 
obituary citation online still had to contact the library to obtain more 
detailed information or the complete text of the obituary. 

Chelsea District Library was lucky enough to have a collection 
that included full-text clippings of obituaries on a significant portion 
of the cards in the Family History Index. Because of this and advances 
in technology, one of the library's primary goals was to make complete 
text available freely online, meaning genealogists and researchers had 
at least the possibility of meeting their research needs without leaving 
their computer or contacting the library at all. To further this goal, the 
library sought to develop a database that was free, simple to use, and 
contained relatively small image files accessible even by those with 
dial-up internet connections. 

Important sources, although they sadly have not been updated in 
recent years, were the book and accompanying website, "Moving 
Theory into Practice," and the Making of Modern Michigan project. 
Moving Theory into Practice: Digital Imaging for Libraries and 
Archives by Anne R. Kenney and Oya Y. Rieger (Mountain View, CA: 
Research Libraries Group, 2000) provided a good grounding in what 
to consider when planning a digitization project, as well as minimum 
standards for use and long-term preservation. An online tutorial of 
housed at Cornell University Library (2010) offered step-by-step 
instructions for taking a project successfully through to completion. 
The Making of Modern Michigan was an IMLS-funded joint effort by 
the Michigan State University Library, the Library of Michigan, the 



Digitization on a Dime 45 



Michigan Library Consortium, and others (Michigan State University 
Libraries, 2005). Although its heyday had already passed by the time 
Chelsea embarked project, the website provided helpful background 
information on the structure of digitization projects and, especially, 
scanning equipment that had been vetted. The scanner models in the 
Making of Modern Michigan are no longer produced and libraries may 
have moved to higher minimum DPIs for scanning, but the advice 
offered at both sites remains sound. 

Two more recent resources, available freely online, are BCR's 
CDP Digital Imaging Best Practices, from the Bibliographical Center 
for Research in Colorado (Collaborative Digitization Program, 2008) 
and NISO's IMLS-funded A Framework of Guidance for Building 
Good Digital Collections (National Information Standards 
Organization, 2007), which within the framework includes links to 
many other resources on more specific topics. 

Project scope 

Armed with this information, Chelsea District Library took a look 
at the Family History Index. While the cards had been stored away 
from the light in filing drawers, they also had been used heavily by 
patrons over the years, resulting in some wear and tear. In addition, 
neither the index cards nor the glue were acid free, and newsprint is 
one of the most acidic papers. The oldest cards and the oldest 
clippings dated back to the 1950s, resulting in some that were in very 
delicate condition. Along with the unique and irreplaceable nature of 
the collection, this delicate condition led the library director to decide 
that the cards should be scanned manually in-house rather than sent 
to a contractor who would likely feed them into a scanner, possibly 
resulting in damage. 

This decision meant significant labor for library staff members 
and volunteers. In May 2005, the library applied for a grant from the 
State of Michigan to fund the digitization, which would have allowed 
for the hiring of contractors, but the application was turned down. 
Luckily, the library already had a well-established volunteer services 
program. Lacking any funding beyond $5,000 committed by the 
library, the project manager developed a volunteer-driven plan and 



46 Digitization on a Dime 



made the choice to seek an open source solution for the database, 
resulting in savings on the equipment side. An RFP was opened to 
competitive bidders in the fall of 2005, with selection of a contractor 
and design and testing of the database completed by early 2006. 

During this time, the library also made decisions about the scope 
of the project. While the collection itself was estimated to consist of 
more than 50,000 cards, closer to 25,000 obituaries were 
represented, due to Jones' system of cross-referencing women by 
maiden name. A database eliminated the need for this. The remaining 
set of obituaries came from a number of sources: 

• gravestone transcriptions from Chelsea's three cemeteries and 
several others in neighboring areas 

• notes culled from early histories of the area, as well as 
scrapbooks and other material in the local history collection 

• notes from death notices published in the local newspaper, The 
Chelsea Standard, taken from microfilmed versions of the 
paper, dating from about 1887-1950 

• complete obituaries from The Chelsea Standard, clipped and 
pasted on cards, dating from about 1950 to the present 

• complete obituaries from the newspapers in two nearby cities, 
Ann Arbor and Jackson, clipped and pasted on cards, dating in 
the 1970s and 80s. 

Of this material, the first three could be included in the project 
without further consideration of copyright, as the donation of the 
Family History Index to the library included rights to copying the 
material. The Chelsea Standard, a weekly publication owned by 
Heritage Newspapers, supported the project from the beginning, 
granting copyright release for material originally printed in its pages. 

Unfortunately, the publisher of both the Ann Arbor News and 
Jackson Citizen-Patriot declined to grant copyright permission. The 
library considered this a minor setback, as obituaries from those two 
newspapers covered a span of only about 20 years. The impetus for 
collecting from newspapers in the neighboring cities was that some 
Chelsea area residents chose to publish obituaries only in these 
publications. The number of citizens who fit this scenario, however, 



Digitization on a Dime 47 



was greatly outweighed by the number of obituaries of people who had 
no connection to the library's primary service area at all. The library 
considered entering data from these obituaries but refraining from 
scanning them, which would not have violated copyright law, but 
chose instead to leave the cards for a potential future project and focus 
instead on truly local residents. 

Staffing 

At this project's initiation, Chelsea District Library had about 
twenty employees. The project could not be completed by paid staff, 
and the small budget precluded the hiring of contract labor. 

The library had a well-established and strong volunteer program, 
including a tradition of volunteers working with the Family History 
Index. Started as a volunteer effort in 1932, the library had always had 
strong support from volunteers, and genealogy in general is a topic 
that draws volunteers. Nevertheless, bringing in volunteers double the 
size of the library's own staff would not have been possible without a 
coordinator, who had developed procedures, documentation, and 
processes for intake, training, and evaluation. 

Building on this strong foundation, the library advertised through 
its newsletter, the local newspaper, word-of-mouth, and presentations 
to organizations such as the county genealogical society. The Family 
History Index was a well-used collection and one of the best sources 
for obituaries for the area, so the library was able to to draw in 
volunteers who did not live in the Chelsea district itself. Volunteers 
filed standard application forms, which covered basic contact 
information, times available for work, and special skills. The initial 
group of about three dozen volunteers received training at one of two 
sessions set up in early February 2006, after which additional 
volunteers received training one-on-one or in small groups as they 
signed on. Later, existing volunteers would train new recruits. Over 
the course of the project, nearly 50 volunteers contributed to various 
aspects of the project. 



48 Digitization on a Dime 



Database development 

In considering technology options the library's predominant 
constraints related to both funding and expertise. In 2005, Chelsea 
District Library contracted most of its technology services to the 
library cooperative of which it was part, including website hosting. 
The library itself at the time had only six public computers and about 
a dozen non-networked staff computers, with the single on-site server 
running the public computer time management system. No one on 
staff had knowledge of server administration, leaving staff nervous 
about hosting a server for the digitization project but also open to any 
of a number of configurations. 

Open source software has seen increasing adoption by public 
libraries in recent years. While often referred to as "free," open source 
software is monetarily free only in the sense that to obtain a copy of 
the code requires no exchange of funds. In the truer sense, "free" 
refers to the user's freedom to view and adapt the software, generally 
with an agreement to then share improvements with the larger 
community. Chelsea District Library initially considered both 
proprietary and open source options for this project, as well as both 
in-house and contract solutions. While the librarian hired as project 
manager had extensive experience with Microsoft Access, the 
licensing costs for the accompanying Microsoft SQL server allowing 
multiple simultaneous users made that option prohibitive. An open 
source database based on PHP and MySQL appeared to be a more 
realistic option, with the drawback that no one on staff had the 
requisite familiarity with these programming languages. The library 
decided to solicit requests for proposals to get a better sense of its 
options. Replies to the RFP highlighted the range of options: from a 
$40,000 proposal that involved proprietary software and taking cards 
offsite for more efficient scanning to a $1,500 proposal based on open 
source software and leaving data entry and scanning purely up to the 
library. The library selected the able services of a programmer who 
worked at a nearby library and who recommended the purchase of a 
server and quickly designed and built a PhP/MySQL database meeting 
specifications. While open source was not the initial goal, that such 



Digitization on a Dime 49 



software ended up forming the basis for the database contributed 
greatly to the library's ability to produce a high-functioning, easy-to- 
use database on a limited budget. 

Once the library selected the underlying software, the real work of 
database design began. Priorities were a simple interface on both the 
administrative and public ends; completely web-based access for both 
data entry and retrieval; and the ability to attach multiple images to 
each record. The library pictured a database that would be simple and 
fast for access by users all over the world and one that could be 
expanded to meet larger goals for the local history collection. Starting 
with attached obituary images, the library envisioned eventually 
allowing researchers to submit their own photos, family trees, 
marriage licenses, or other material that could supplement the 
library's own collection to tell the broader story of each person 
represented in the database. Flexibility for growth of both the size and 
scope of the collection was important. 

These ambitious plans remained in the future. In the near term, 
the library had to balance providing extensive access to the obituaries 
with completing the work within a relatively brief time frame. The 
poor quality of the newsprint and cards meant doing optical character 
recognition (OCR) was not realistic. Thus, while researchers would be 
able to view an image of the complete obituary, searching would be 
limited to data entered by volunteers. This meant maximizing the 
number of access points was ideal; at the same time, too many access 
points could slow work to a crawl. In the end, after consultation with 
genealogical researchers, the library chose the following fields: 

• first, middle, last, and maiden names of subject in separate 
fields 

• first and last names in one field for: mother, father, spouse(s), 
children 

• date and place (city, state) of birth and death 

• cemetery and funeral home 

• metadata covering obituary source and date and source of 
digital record 



50 Digitization on a Dime 



The database allowed for multiple spouses and children. It also 
included a notes field which, at the early stages, was left blank but 
proved to be invaluable for later expansion. 

The database would be accessed via the library's website by users 
through either a basic or advanced search. Upon visiting the Family 
History Index Online, users see a basic search screen for the name 
field, which searches all name-related fields. This is often enough to 
get users to the obituary they need. 

Users also have access to an advanced search screen, which 
searches first, last, and maiden name as separate fields. It also allows 
access via date and place of birth or death, cemetery, and funeral 
home, making it useful to those who may not be searching for a 
specific person but for more general historical information. Data 
typed into any of the fields on the basic or advanced search screens 
will also search the notes field, which may contain additional data 
from various sources. 

The designer brought another feature to the database that would 
set the Family History Index Online apart from similar projects at 
other libraries, a hyperlinking feature that makes Chelsea's project 
unique in allowing researchers to jump from record to record, 
following the obituaries of family members represented in the 
database. If the parent, child, or spouse of an obituary subject is 
represented in the database, his or her name will appear as an active 
link. Clicking on this link takes the user to that person's obituary 
record. In this way, researchers may discover family connections they 
did not know existed and be better able to visualize how families relate 
to each other. Rather than noting the names of relatives, backing out 
to the initial search screen, and starting a new search, database users 
can simply hop from one relevant record to the next. 

Once the database design was complete, both staff and an initial 
group of volunteers participated in testing. Volunteers tested for ease 
of use from the administrative and public perspectives as well as for 
how the database would meet the needs of genealogy researchers. The 
library was lucky to have a number of experienced genealogical 
researchers among its volunteer corps. They provided invaluable 



Digitization on a Dime 51 



feedback throughout the project. The database designer and project 
manager worked together to tweak the database in late 2005 and early 
2006, leading up to training of volunteers and beginning of data entry 
work. 

Timeline 

• April 2005 - Chelsea District Library creates a librarian position 
with duties including digitization of the Family History Index 

• October 2005 - Database development begins 

• January 2006 - Database development and testing completed 

• February 2006 - Volunteer training and data entry work begins 

• May 2006 - Scanning of obituary cards begins 

• October 2006 - Data entry of 15,000 records completed/ library 
closes to move to new building 

• January 2007 - Scanning resumes; proofreading and database 
updates continue 

• June 2007 - Final image attached to database 

• October 2007 - Family History Index Online released to public; 
timed to coincide with Family History Month in Michigan 

Digitization process 

The library had two old PCs available for use by volunteers in a 
back office, a setup that proved beneficial to the project, as volunteers 
found they could often focus better on the mundane task of data entry 
while working in pairs. Because the database was entirely web-based, 
no software installation was required. Volunteers also used each other 
as resources to answer questions such as the interpretation of unclear 
wording or how to enter data in a particular field. Each volunteer had 
committed to working two hours per week for a period of at least three 
months, in an attempt to minimize the amount of retraining that 
would need to be done. Most volunteers stayed much longer, seeing 
the project through to completion, and many also worked multiple 
shifts each week. Through the volunteer program, these workers were 
able to sign up for shifts during all hours the library was open, 
including evenings and weekends, maximizing the number of people 
who could be involved. 



52 Digitization on a Dime 



As the project progressed, it became clear both that some 
volunteers wanted to participate but lacked interest or ability in data 
entry and also that volunteer tasks existed beyond what the library 
had originally envisioned. This led to some refocusing of efforts before 
a final workflow developed. In particular, a pair of volunteers took on 
the task of taking cards from the filing cabinet and sorting out those 
that need digitization. They stored the sorted cards in a box and 
transported them to the office where volunteers were engaged in data 
entry. Cards not in use remained in the files, which prominently 
displayed signs explaining the project. Volunteers doing data entry 
took the sorted cards and entered them into the database. Cards that 
brought up questions went into a separate pile for review by the 
librarian. The rest went into a "completed" file and moved on to step 
2, proofreading. Two volunteers with especially good eyes for detail, as 
well as genealogical research experience, were recruited as 
proofreaders, tasked with checking every fifth card. While it would 
have been ideal to have a second set of eyes on every single database 
record, this was not realistic, and conferral with professional 
archivists confirmed that a 20 percent rate was more than sufficient. 

After proofreading, cards moved on to scanning. Scanning 
represented another challenge, in part because the library had only 
one scanner which was also used for other purposes. In addition, 
while the scanning software that accompanied the purchase HP 
Scanjet 5500c was relatively simple, not all volunteers felt they had 
the requisite level of computer skills. As a result, scanning did not 
begin until about three months after data entry and was handled by a 
subset of about 10 volunteers. Volunteers scanned cards at 300 dpi 
and saved them in the archival standard TIF format. Because they 
were on black-and-white newsprint and newsprint already has a very 
low resolution, a higher resolution would not provide any benefits. 
These archival copies of the cards have been retained in separate, 
backed up files so that the original cards should never require 
rescanning. The entire database, including these archival images, was 
set to copy to a tape drive, with the tape changed daily. After one 
week, tapes were reused for new backup copies, leaving the library 
with multiple recent backups for added security. 



Digitization on a Dime 53 



At this point, the cards were refiled by a volunteer, returning 
access to the public who still relied on the paper file for research. The 
digital images then underwent additional processing. The TIF images 
were converted, using Adobe Photoshop Elements into compressed 
JPEG images with small file sizes. In general, the image of each card 
posted to the database was no more than 100 Kb. It would download 
quickly even with dial-up internet connections. Volunteers manually 
attached these images to each database record, completing the cycle. 

Data entry work progressed remarkably quickly, with volunteers 
putting in close to 2,000 hours between February and October 2006. 
In October 2006, the project went on hiatus while the library packed 
up its temporary quarters and moved into a new facility. Volunteers 
completed data entry for the final card just before the move 
commenced, adding more than 15,000 records into the database in 
eight months. Between half and two-thirds of the scanning had also 
been completed at this point. The move to the new facility caused 
significant delays in the project due to other priorities for the library's 
technology staff. Scanning resumed in February 2007, however, and 
the final image was attached to the database in June. The library used 
the next few months to continue proofreading and to test the 
robustness of the database, releasing it to the public in conjunction 
with Michigan's Family History Month in October 2007. 

Results 

Chelsea District Library's Family History Index digitization 
project was a major success. Through a combination of creativity, 
open source software, and volunteer contributions, the library 
produced a highly usable online database providing full access to more 
than 15,000 obituaries for less than $5,000, completing work in about 
15 months. It demonstrates that even at an institution with little staff 
and little specific expertise, a project can be developed to meet both 
the community's needs and the standards set by the library and 
archives world. The community involvement and grassroots nature of 
this project made it truly special for those who participated. It drew 
attention to the library, increased the volunteer corps, and gave 



54 Digitization on a Dime 



community members a sense of ownership for a key piece of the 
library's collection. 

The flexibility with which the database was designed has proven 
vital to its success. Once done with the initial work of populating the 
database, the library turned its attention to improvements. One goal 
was to provide full-text access to all obituaries, even those from 
newspapers old enough that the only existing copies were on 
microfilm. In early 2008, the Friends of the Library applied for and 
received grant funding to help purchase a digital microfilm machine, 
which volunteers are currently using to scan obituaries from the 
Chelsea Standard going as far back as copies survive, into the late 
1800s. These digital images are then either added to existing database 
records or used to supplement the database. 

There also remained the problem of providing access to obituaries 
of local residents that were printed in neighboring newspapers. To this 
end, the library looked to its partnerships with local businesses. 
Chelsea is represented by two local funeral homes, and directors of 
both proved willing to provide the library with access to their files. 
When the funeral homes submit obituaries to any area newspaper, 
they also send an electronic copy, including photo where available, to 
the library. These partnerships have allowed the library to enrich the 
database with full-text access and color photos for more recent 
obituaries. The text that is sent electronically is pasted into the notes 
field, and images are higher quality than those ultimately printed in 
the newspaper. Along with this material, if the final obituary is printed 
in the Chelsea Standard, the scanned newspaper clipping is attached. 
The funeral homes have also indicated that they have computerized 
files going back a number of years, and the library is investigating the 
possibility of further enhancing the database with this material. 

As another extension, the library returned to its initial contractor 
in 2008 to develop a database on the same platform to house the 
library's local history collection, which had been brought out of 
storage and organized only in 2007. Much of this material was even 
more valuable and unique than the obituaries, leaving the library with 
a strong desire to have it made accessible to the public primarily, if 



Digitization on a Dime 55 



not exclusively, in digital format. Scanning of material and database 
development have continued in 2009. While working on this and 
other local history initiatives, including a series of oral history 
projects, the library has found additional material to enhance the 
obituary database records of members of Chelsea's founding and 
prominent families. 

Aside from the primary lesson that ambitious digitization projects 
are not beyond the reach of even very small libraries, the Family 
History Index digitization project provided additional lessons that 
apply to similar projects and beyond: 

• The existence of a well-setup volunteer program allows a library to 
think much bigger than would otherwise be practical. People are 
out there in all communities who have the time, expertise, and 
interest to contribute. 

• Planning is good but flexibility is essential. Projects may stray 
somewhat from their original vision or carefully thought out 
procedures, but that isn't necessarily bad. Being open to new 
opportunities and listening to workers and users can ultimately 
make a project much richer. 

• Open source doesn't have to be terrifying. In fact, open source 
solutions are generally very stable, as the Family History Index 
Online has proven to be. Aside from minor software upgrades, the 
server has provided consistent access to the database for more 
than two years with virtually no staff intervention required. 

• There's nothing wrong with thinking big, but it doesn't hurt to be 
realistic. Ambition leads to projects being even more successful 
than initially imagined. That said, a realistic assessment of aspects 
such as which tasks could be handled in-house (project 
management) and which should be contracted out (database 
design) prevented later stumbles. 

• Partnerships enhance any library activity. In this case, a good 
working relationship with the local newspaper eased the process of 
gaining copyright access, and new partnerships with funeral 
homes have enhanced both the database and the library's 
reputation in the larger community. In other situations, 



56 Digitization on a Dime 



partnerships could be used to gain access to services or material 
the library cannot pay for or obtain on its own. 

Small libraries often represent their communities' best hope for 
preserving local history. Preserving this history, even with little or no 
budget, is within reach if libraries combine their expertise in 
information management with technology decisions geared toward 
simplicity and a lack of hesitation in taking advantage of the 
knowledge and goodwill in their communities. Digitization projects 
not only preserve the past but provide an opportunity for greater 
community involvement, partnerships, and identification of the 
library as a key to the community's overall health. 

References 

Chelsea District Library. (2009). Chelsea District Library Family 
History Index Online. Retrieved on Jun 30, 2010 from 
http://fh.chelsea.lib.mi.us/ 

Collaborative Digitization Program. (2008). Western States digital 
imaging best practices, version 2.0. Retrieved on March 31, 2010 
from http://www.bcr.org/dps/cdp/best/wsdibp_vi.pdf 

Michigan State University Libraries. (2005). The making of modern 
Michigan: Digitizing Michigan's hidden past. Retrieved March 15, 
2010, from http://mmm.lib.msu.edu/. 

Cornell University Library. (2010). Moving theory into practice 
digital imaging tutorial. Retrieved March 15, 2010, from 
http : / / www. library. Cornell . edu/ preservation/ tutorial 

National Information Standards Organization. (2007). A framework 
of guidance for building good digital collections. Retrieved March 
15, 2010, http://framework.niso.org/ 



Building the ALBA Digital Library 



Jill Annitto (Archivist) 



Abstract 

This chapter serves as a case study of how a professional digital library 
can be successfully built with a small staff and budget. It discusses the 
planning and experiments with beta versions of the Abraham Lincoln 
Brigade Archives (ALBA) Digital Library, the final version of which is 
available on ALBA's website, through Metropolitan New York Library 
Council's (METRO) Digital Metro New York program, and OCLC 
WorldCat. The sensitive issues of digitizing another institution's 
collection while maintaining ownership of the final product are also 
explored. 

Keywords: Copyright, CONTENTdm, Database, Digital archive, 
Electronic classroom, Educational resources, Funding, Indexing, 
Planning, Ownership issues, Small budgets 



Introduction 

In 1979, recognizing the vital importance of their radical history, and 
the need to collect writings, letters, photographs, oral histories and 
artifacts that would preserve their story, the Veterans of the Abraham 
Lincoln Brigade, the American volunteers who fought with Republican 
forces against Generalissimo Francisco Franco during the Spanish 
Civil War (1936-39), formed the Abraham Lincoln Brigade Archives 
(ALBA). Today ALBA lends its name to a major archive at New York 
University's (NYU) Tamiment Library and independently supports 
cultural and educational activities related to the war. The ALBA 



57 



58 Building the ALBA Digital Library 



collections are the most requested at the Tamiment Library. ALBA's 
relationship with NYU is non-traditional, which makes the ALBA 
Digital Library unique. NYU owns and maintains the Archives and 
hosts many of ALBA's programs, yet the ALBA name gives a different 
impression. 

This chapter explores the challenges of digitizing a collection that 
is owned by another institution and how to overcome working with a 
limited budget, dated technology, and minimal staff to produce a 
professional digital resource. The ALBA Digital Library (Abraham 
Lincoln Brigade Archives, 2008) evolved from an ineffective form- 
based website tool to a fully indexed resource using CONTENTdm in a 
matter of months. 

ALBA Goes Digital 

Until a full-time Executive Director was hired in 2007, ALBA was 
initially run by a group of volunteers followed by a string of part-time 
administrators. The ALBA Board of Governors hired me as the 
Assistant Director in May 2008 just as they began reconsidering their 
outdated website, a major step for the small organization. 

By June 2008 the Executive Director had resigned, leaving me as 
ALBA's only employee. It provided me with a great opportunity to 
help redesign the website alongside a subcommittee of the Board of 
Governors. The Web Committee, consisting of four history professors 
living across the United States, set a timeline of four months for the 
site's overhaul, with an anticipated launch date of early October. The 
Board wanted to stake a claim as the premier electronic resource for 
information on the Spanish Civil War. 

The Impetus for a Digital Library 

In July 2008 ALBA launched its first annual ALBA Teachers Summer 
Institute at NYU. The Institute hosts teachers from New York City 
public schools and exposes them to the history, art, and politics of the 
Spanish Civil War. Part of this immersion includes a trip to NYU's 
Tamiment Library to view the Brigade's Archives where teachers 



Building the ALBA Digital Library 59 



receive an overview by the collection's archivists as well as professors 
from various departments at NYU. 

After this initial success, ALBA decided to sponsor two more 
Summer Institutes (as well as year-round professional development 
seminars) in Tampa and San Francisco in 2009. The new settings 
precluded a site visit to the Archives so we needed to find a way to 
bring the Archives to the teachers. 

The solution to this dilemma was the creation of a digital library 
or archive to be launched in conjunction with the new website. The 
website redesign was going to make ample use of ALBAs collection at 
NYU's Tamiment Library anyway, slowly integrating digital 
collections of letters and photos, and eventually including video clips 
and oral histories. These primary resources would strengthen existing 
lesson plans and other educational modules once the site was 
completed and allow for a more dynamic classroom experience. Since 
the site was already being designed, we thought it would be a great 
opportunity to have a collections database created for ALBA by our 
website designers. 

Defining the Digital Library 

In my experience archives or library staff working in conjunction with 
a representative from the information technology department does 
most digital library planning. At ALBA, I was the only staff member 
and every decision required consensus of the Web Committee, busy 
professionals with jobs and other responsibilities. Every time an issue 
arose it could take nearly a week to be resolved. This is a problem that 
small museums with very active boards will be familiar with, 
particularly when board members have little time, experience or 
interest in the digital project at hand. 

We quickly ran into a problem of defining the digital library; the 
Web Committee believed everything created by ALBA was archival. I 
was in favor of a more traditional, primary source, collections-based 
digital library with a thesaurus and cataloging guidelines. My idea for 
ALBA's digital library was to create a system that would recall only 
primary sources from the war itself. The digital library would be its 



60 Building the ALBA Digital Library 

own entity containing items that are separate from all other files 
posted to the website. 

After several weeks of negotiations via email and conference calls, 
the Web Committee decided the best option would be to include all 
online documents in the digital library. The Web Committee charged 
the website designers with creating a web-based, simple form-based 
recall system as part of their contract. 

The Beta Versions 

Albita 

Named Albita (or 'little ALBA') this resource was first launched in 
November 2008 along with the debut of the new website. It included 
the organization's quarterly newsletter, book reviews, and transcribed 
letters, not primary archival sources. Uploaded items were listed in 
random order rather than alphabetically or by date. In the end the 
design did not conform to the standards of information professionals; 
Albita had become a "junk drawer" of every single file uploaded to the 
website. 

Document Library 

Even renaming it and reconfiguring the display, the Document 
Library was not robust enough. To recall specific items, the user had 
to consciously use certain keyword strings, e.g. "George Watt Prize 
Winner," when retrieving items. It did not include a thesaurus and the 
plain-text description field did not allow for paragraph breaks. These 
descriptions were displayed as a solid block of text. 

Media Library 

Simultaneously added to the site was the Media Library. This 
database appears three pages deep within our Resources tab. It is the 
only way to access images (from archival photographs to logos of 
partner organizations) without using the ALBA Digital Library or 
slogging through the results from a general site search. Unfortunately 
because it is not linked to any other page, it is not often accessed. The 



Building the ALBA Digital Library 61 



plain-text description field is displayed as a block of text and does not 
allow indexing. 

Once the Web Committee saw Albita, the Document Library, and 
the Media Library in action the problems became apparent. We 
considered going back to the drawing board with our website 
designers, to create a thesaurus or a more sophisticated database 
system. Unfortunately, our original contract did not cover building a 
true digital library; the cost would have been prohibitively expensive. 

I explained the possibilities of a professional digital library: 
adding our records to a consortium's collection, making them widely 
available through WorldCat, including them in an NYU catalog at the 
item level. It was difficult to convince the Web Committee that we still 
had affordable, professional options that would blend seamlessly into 
the new website. 

Building a Better Mousetrap 

As the only person on staff who could design and implement the 
digital library, I had to balance time spent on digitization with my 
other duties, including grant writing, public programming, and 
administrative issues. 

The first problem for ALBA to overcome was the lack of any type 
of digitization equipment. The nature of ALBA's work and budget did 
not necessitate having a full flatbed scanner or laptop on hand. 

Another major concern for ALBA was the high cost of software. A 
rough estimate for a full software package ran to the thousands of 
dollars. We didn't intend to scan the entire collection, just highlights; 
for our purposes what we needed was something that would allow us 
to scan part of the collection but offer maximum exposure. 

In January 2009 I became aware of the Metropolitan New York 
Library Council's (METRO) Digital New York program. NYU's status 
as an institutional member of METRO allowed ALBA to benefit from 
the digitization program, a partnership between METRO and OCLC 
and CONTENTdm. As a separate organization, ALBA's much smaller 
annual budget resulted in a nominal $200 annual fee, paid to METRO 



62 Building the ALBA Digital Library 



for access to the CONTENTdm desktop module and the upload of 500 
discrete items. With software issues out of the way, we could focus on 
permissions and access to the collection. 

Since ALBA's intention was to work with a collection that is 
owned by another large institution, we had to be very thorough and 
diplomatic about the project. We presented a sample record that 
would name NYU and Tamiment Library as the copyright holders, as 
well as instructions on obtaining permission to use the images. We 
assured NYU that the digital library pages on ALBAs website would 
include the same information. A distinction is made on both the 
website and in each record that the ALBA Digital Library is published 
by ALBA while the copyright is held by NYU. After several weeks of 
negotiation, we were free to move ahead with the project. 

With software and permissions settled we were left with the issue 
of hardware. ALBA Board Member and NYU Professor of Spanish and 
Portuguese James Fernandez offered his laptop and Epson flatbed 
scanner for the duration of the project. After several weeks of further 
negotiation in order to gain access to the collections, Michael Nash, 
ALBA Board Member and Director of Tamiment Library, allowed us 
to scan the items ourselves, free of charge. 

Selection, Policies, and Standards 

In general there is a lack of digitized archival resources available on 
the Spanish Civil War. The closest to any kind of digital information 
on the subject is through Spartacus Educational (Simkin, 1997), an 
online British encyclopedia dedicated to educating students on 
history, with a significant amount of information on progressive 
history. But even this site is lacking in primary source materials. 

Besides the Abraham Lincoln Brigade Archives, there are other 
Spanish Civil War archives in the United States: the Southworth 
Collection at the University of California at San Diego (see University 
of California, San Diego, n.d.) and the Spanish Civil War Collection at 
the University of Illinois - Urbana Champaign (see University of 
Illinois at Urbana-Champaign, n.d.) Still, the Abraham Lincoln 



Building the ALBA Digital Library 63 



Brigade Archives at NYU is the largest collection of American 
volunteers' archives in the US. 

Our goal was to put forth a curated collection of the most 
compelling and historically significant items in the Archives; the 
original plan was to digitize 500 items, but the realities of time and 
staff restraints reduced this number to an initial 150 items. These 
items included postcards, letters, newspaper articles, a multi-lingual 
newsletter, and a telegram from Ernest Hemingway. 

Dr. Fernandez performed the selection of four collections based 
on the following criteria: size (only one manuscript box each), 
condition (stable enough for handling and scanning), and variety 
(each collection included letters from a variety of people). These same 
collections had been used in the Teachers Summer Institute and they 
proved to be popular with the teachers. 

When it came to setting the scanning standards I turned to the 
University of Wisconsin-Milwaukee Libraries' Digital Collections (see 
University of Wisconsin-Milwaukee, 2006) pages, namely the 
Transportation Around the World, 1911-1993 collection that I worked 
on as a graduate student there in 2003. The collection was built using 
CONTENTdm, then in its infancy. I followed UWM's digitization 
standards and policies as well as their formula for long-term archival 
storage as a basis for the ALBA Digital Library. 

Digitization Standards 

All items were scanned in full-color at a resolution of 600 dpi 
using an Epson flatbed scanner. The items were saved as the highest 
quality TIFF files and stored on an external hard drive. We refer to 
these TIFF files as the Archival Images as they are used only to create 
access images (PDFs or JPEGs) and are otherwise not accessed. 

Access Images were created for web delivery, in the form of 
thumbnails on the ALBA site and for display within the CONTENTdm 
records, as well as for everyday use. The letters in the collection do not 
have OCR capabilities nor are the PDFs searchable. These were saved 
on both DVDs and the external hard drive. Because the METRO 
contract only covered 500 discrete items, we used Photoshop to stitch 



64 Building the ALBA Digital Library 

the JPEGs together to create a single PDF file for multiple-page items. 
Not only did this save room in CONTENTdm, it was also useful to 
have multiple-page documents combined for reproduction purposes 
and to better keep track of collections. 

All of this information is posted on the Digitization page of the 
ALBA Digital Library section. Sharing this information shows 
researchers (and potential donors) that the digital library adheres to 
archival standards while also helping other librarians and archivists 
plan their projects. 

To access the images it is important to label them properly. Some 
scanners and scanning programs assign numbers that may or may not 
be useful. Auto-numbering systems can cause problems if the items 
are not easily identified by sight (letters, manuscripts). I named the 
files according to the manuscript collection followed by a number that 
corresponded to the letter's order in the folder, followed by a decimal 
that corresponded to the page number. For example, Lardner.1.4 
would be the fourth page of the first letter of the Lardner Collection. 
The stitched PDF file would read Lardner. 1. 

Indexing 

The documents were indexed using Dublin Core metadata 
(Dublin Core Metadata Initiative, 2010) which are standard in 
CONTENTdm, including: Creator, Date, Identifier, Type, Source, 
Description, Format, Coverage, References, Relation, Language, 
Publisher, Rights, and ALBA Reference Number. 

Included in CONTENTdm is a default thesaurus (Thesaurus for 
Graphic Materials) for the Identifier field. Similarly, a default 
thesaurus provided the terms for geographic location in the Coverage 
field. Those fields for which the thesaurus was unique to the Abraham 
Lincoln Brigade Archives, I built a new thesaurus. These fields were: 
Type (e.g. text, image), Format (e.g. paper + size in centimeters), 
Identifier (e.g. envelope, autograph letter signed, typed letter), 
Publisher, Rights, and Creator. I added new Creators as they came up 
in each collection; some collections had as many creators as there 
were letters. In addition, for collections that had an online finding aid 
at NYU I included a link to that page. 



Building the ALBA Digital Library 65 



In all, the ALBA Digital Library took four months for 150 items to 
be scanned, stitched, catalogued, and uploaded. I generally spent 
about fifteen minutes cataloguing each letter, sometimes longer 
depending on the length of the document. I worked on the digital 
library an average of eight hours per week with some weeks going by 
without any work at all. As collections were completed I added 
information to the ALBA Digital Library page and continued to 
redesign the website as necessary. 

Final Product and Reception 

The final product can be found at http://www.alba-valb.org 
/resources/digital-library. The Teachers Institute Alumni find the 
resource easy to use and helpful in the classroom, and ALBA even 
used the scanned collections to create facsimile copies of the Archives 
for the Tampa and San Francisco seminars. 

As of publication, the digital library is available as a discrete 
collection on Digital Metro New York (http://cdmi2840i.cdmhost. 
com/cdm4/search.php), the digital program of the Metropolitan New 
York Library Council, the New York Heritage Digital Collections (New 
York Heritage, 2008) website, as well as at the item level on OCLC 
WorldCat (http://www.worldcat.org). 

Seeing a Digital Project from Beginning to End 

Planning is Key 

Work backwards and set aside a day or two to really think about 
what you want to see in the final product. The small team (just myself 
most of the time) helped keep bureaucracy to a minimum. Remember: 
it is cheaper to do it right the first time. 

Unofficial market research 

I asked librarian and archivist friends what they liked to see in 
digital libraries and to give me advice on moving forward. I also 
reflected on what struck me in online collections, both the positive 
and the negative. 



66 Building the ALBA Digital Library 



Tread Lightly 

Occasionally it was necessary to receive approvals and it was 
important to remember that when dealing with large institutions there 
will be politics. This exists everywhere and I learned not to take it 
personally. Many of these issues were deep-seated and existed long 
before I joined the organization. I also learned that having a board 
member installed in a specific department or company does not 
guarantee easy access or donated materials. 

Within my own organization, some board members did not 
understand the potential of planned, professional digital libraries, or 
that one person could undertake such a project. Price was also an 
issue; it was only after Albita and the Document Library failed was 
CONTENTdm considered and accepted as an amazing deal. In the 
end, waiting for the Web Committee members to come around in their 
own time was the best plan of action for this project. 

Push the PR 

As I completed each collection I sent information about the digital 
library everywhere: from Facebook and Archivists' Roundtable of 
Metropolitan New York to ALBA's listserv, e-news, quarterly 
newsletter, and fundraising appeals. While this publicity was mainly 
sent to people within the ALBA network, it also garnered the attention 
of local archivists and library students interested in doing small digital 
projects on limited resources. 

References 

Abraham Lincoln Brigade Archives. (2008). ALBA Digital Library. 
Retrieved March 30, 2010, from http://www.alba-valb.org/ 
resources / digital-library 

Dublin Core metadata initiative. (2010). Retrieved March 30, 2010, 
from http:/ / dublincore.org 

New York Heritage. (2008). Retrieved March 30, 2010, from 
http://www.newyorkheritage.org 



Building the ALBA Digital Library 67 



Simkin, J. (1997). Spartacus educational. Retrieved March 30, 2010, 
from http://www.spartacus.schoolnet.co.uk/Spanish-Civil- 
War.htm 

University of California, San Diego, (n.d.). Southworth Spanish Civil 
War Collection. Retrieved March 30, 2010, from UC San Diego 
Libraries website, http://libraries.ucsd.edu/locations/mscl/ 
collections/southworth-spanish-civil-war-collection.html 

University of Illinois at Urbana-Champaign. (n.d.). Spanish Civil War 
Collection. Retrieved March 30, 2010, from University of Illinois at 
Urbana-Champaign Rare Book and Manuscript Library website, 
http:/ /www.library.illinois.edu/ rbx/ SCWPeople.htm 

University of Wisconsin-Milwaukee. (2006). Transportation around 
the World, 1911-1993. Retrieved March 30, 2010, from Digital 
Collections -Transportation around the World, 1911-1993 website, 
http : / / www4 . uwm.edu /libraries/ digilib / transport / index, cf m 



Digitization and Access of Louisiana Oral 
Histories: One Oral History Center's 
Experience in the Digital Realm 

Gina R. Costello (Louisiana State University Libraries) 



Abstract 

The Louisiana State University (LSU) Libraries Center for Oral 
History began an effort to digitize at risk and high demand collections 
in 2007. The Center acquired digitization equipment, server space, 
and collaborated with the Libraries Special Collections Digital Services 
librarian to offer digitized oral histories online via the statewide 
Louisiana Digital Library (LDL). This paper details the history of the 
ongoing development of a digitization program for oral history 
materials using two staff members and limited resources. Decisions 
about what materials to digitize and how, equipment and software, 
and issues with access and preservation will be discussed. 

Keywords: Audio digitization standards, CONTENTdm, Digitizing 
audio, Digitization equipment, Digital library, Digitization workflow, 
Oral history, Oral history interviews. 



Introduction 

The Louisiana State University (LSU) Libraries T. Harry Williams 
Center for Oral History began to digitize at risk and high demand 
collections in late 2007. Planning for the systematic digitization of the 
primarily analog collection began a year prior to any digitization 
efforts. The Center sought advice from an expert in the field, acquired 



68 



Digitization and Access of Louisiana Oral Histories 69 



digitization equipment and server space, hired a full time employee to 
manage digitization, and collaborated with the Libraries Special 
Collections Digital Services Librarian to offer digitized oral histories 
online via the statewide Louisiana Digital Library (LDL). 

The Center staff and the Digital Services Librarian have 
prioritized collections for digitization based on fragility or patron 
demand, made decisions about organization and access of the audio 
materials for the public, and addressed copyright issues. Only a small 
number of oral history collections have been added to the LDL, 
although over 700 hours of tape have been digitized so far. 

This paper details the history of the ongoing development of a 
digitization program for oral history materials with one full time staff 
person and partial effort from another staff member. Decisions about 
what materials to digitize and how, equipment and software, and 
issues with access and preservation will be discussed. Results of the 
digitization and online access efforts have been mixed, but may serve 
as an example for oral history programs wishing to develop a more 
programmatic approach to digitization. 

Center History and Description 

The T. Harry Williams Center for Oral History at LSU Libraries 
Special Collections documents the social, political, and cultural history 
of LSU and the state of Louisiana by conducting, collecting, 
preserving, and making available to the public oral history interviews 
of folk artists, war veterans, governors, congressmen, state and local 
officials, civil rights activists, and other historically prominent figures 
in Louisiana. The Center maintains over 4,000 hours of tape-recorded 
interviews. The three person staff and a number of student workers 
transcribe, index, and deposit oral history interviews for archival 
storage at LSU Libraries Special Collections. 

The Center, opened in 1991, is named after a man who helped 
legitimize the field of oral history. Dr. T. Harry Williams, a popular 
and acclaimed southern history professor at LSU spent more than ten 
years researching the biography, Huey Long. Published in 1969, this 
Pulitzer Prize and National Book Award winning book drew upon 



70 Digitization and Access of Louisiana Oral Histories 

Williams' tape-recorded interviews with nearly 300 individuals. 
Williams used a 30 pound Webster Electric Ekotape reel-to-reel tape 
recorder to capture the interviews. 

The primary mission of the Center is to document the history of 
LSU. Since the history of the state and university are closely 
intertwined, many broader Louisiana subjects are documented as 
well. Public outreach through training workshops, consultations, and 
collaborations with individual researchers, community groups, 
classes, and institutions, enhance oral history collections throughout 
the state. Often, the collections are donated to LSU Libraries for 
preservation and public access. In many cases copies are provided to 
libraries, schools, museums, providing access for members of the 
communities in which the oral histories were collected. 

The Center differs from some oral history centers in its 
commitment to providing fully edited transcriptions of all recorded 
interviews. Barring any restrictions placed on interviews by the 
interviewee or interviewer, the audio and a full transcription are made 
available to scholars and the general public. Because of the large 
volume of interviews that are collected each year, the Center 
maintains a backlog of interviews that are not fully processed (i.e., 
digitized if applicable, transcribed, audited, and cataloged). Interviews 
are organized into more than 40 different series, including Civil 
Rights, Military History, and Political History. 

The Center Director has taken a more programmatic rather than 
project-based approach to the digitization of the collected oral 
histories. To ensure that preservation issues are addressed and 
collection access is a top priority, the Director employs a full time 
sound technician/webmaster at the Center. Center staff also works 
with the Special Collections Digital Services Librarian to mount oral 
history collections to the Louisiana Digital Library (LDL) 
(http : / / www.louisianadigitallibrary. org) . 

The Center makes available materials that are not restricted by 
the interviewee or interviewer. Interviews are digitized on demand for 
patrons, for preservation purposes, and for public access on the LDL. 
Prior to the acquisition of digitization equipment, patron requested 



Digitization and Access of Louisiana Oral Histories 71 



copies were recorded from cassette tape to cassette tape. Now 
materials are delivered to patrons via CD unless a cassette tape is 
requested. Copies are provided for a fee to patrons, although a small 
number of oral histories maintained by the Center are available for 
listening free online in the LDL. Center staff generally digitize fewer 
than five interviews per month for patron requests. 

The funding for the Center is a mix of Libraries monies and 
endowment funding. The Libraries pays the salaries of the Director 
and two full time employees. Student workers' pay, a portion of travel 
money, and some supplies are also paid for by the university. The 
Libraries purchases and provides support for computers for the 
Center staff and student workers. Endowment funds cover most travel 
expenses, the majority of the equipment (specifically the field 
recorders, digitization station, software, fax machine, scanner), any 
Graduate Assistantships, additional student workers, and the majority 
of the transient workers' (e.g., professional interviewers, transcribers, 
editors) wages 

Early Forays in Digital Access 

One of the earliest digital projects the Center was involved with was a 
pilot project to digitize oral histories that are part of the University 
History Series sub-series, Integration and the African American 
Experience at LSU. The sub-series contains interviews with black 
students, faculty, and administrators at LSU during integration (1950- 
1970), plus interviews with lawyers and their clients who were 
involved in key lawsuits, as well as politicians and others who were 
vocal opponents or supporters of integration. The resulting digital 
collection, named "Integration and the Black Experience at LSU" 
(2003) contains audio files and transcriptions of three individuals 
interviewed between 1985 and 1998. 

This legacy digital collection is scheduled to be revamped soon. 
The ".rm" or ".ram" audio files are available for listening only in 
RealPlayer and must be downloaded to the listener's computer before 
playing. The digital files were created more than eight years ago, so 
the sound quality could be improved and the information about 



72 Digitization and Access of Louisiana Oral Histories 

equipment and digitization method has been lost. The analog tapes 
will be re-digitized and optimized using current technologies. 

Between 2001 and 2005, the Center utilized the skills of their part 
time webmaster and other staff members to create several online 
exhibitions and presentations (T. Harry Williams Center for Oral 
History Exhibits and Presentations, 2009) using readily available 
software and tools: simple HTML, PowerPoint, and Windows Movie 
Maker. Notable among these is the digital exhibition, "Baton Rouge 
Bus Boycott of 1953. A Recaptured Past" (2009) which includes a 
background and chronology of the event complete with photographs 
and audio excerpts. "Leaving Vietnam" is a nine minute presentation 
of audio clips from the Americans in Vietnam collection, featuring 
stories of escape from three Vietnamese refugees who immigrated to 
Louisiana around 1975 while fleeing Communist takeover. The 
presentation debuted at the 2005 Oral History Association annual 
conference and is currently available on YouTube, where it has been 
viewed over 6,000 times. Two other presentations were also mounted 
on YouTube to provide ease of access. 

Center staff also began digitizing oral history transcriptions that 
were only available in paper format in 2004. They had some success 
using a HP Scanjet 5590 document feed scanner and an early version 
of Readiris optical character recognition (OCR) software. The software 
was lost, and the Libraries Systems department replaced it with 
Readiris Pro 11. Subsequent digitization efforts have been stymied by 
problems getting good readable OCR text, so the project has been put 
on hold. Student workers often are tasked with re-keying 
transcriptions. 

In 2007 the Center Director, with the help of the LSU Libraries 
Special Collections Exhibitions Coordinator, curated a physical 
exhibition called "Have you Heard?: The Past in First Person from the 
T. Harry Williams Center for Oral History". The extensive exhibition 
contained ephemera and narrative relating to more than a dozen oral 
history collections. The Libraries provided two "listening stations", 
computers loaded with web-based presentations in the exhibit hall. In 
addition, exhibit-goers could check out MP3 players with pre- 



Digitization and Access of Louisiana Oral Histories 73 



recorded narration of the exhibition contents and snippets of oral 
history interviews. These digital offerings were made available with 
little cost using spare computers and a staff member as the voice of 
the narrator. No previous Libraries exhibition had employed 
technology in these ways. The Center Director counts the exhibition a 
success, as it led to a few collection development opportunities and 
awareness of the Center and its mission. 

Digitization Station 

After attending a digitization workshop at the Oral History 
Association annual conference in 2006, the Center Director decided 
that the systematic digitization of at risk and high demand analog 
collections should become a central focus for the Center. With the idea 
of "going digital" but with little research in hand they initially 
purchased two standalone analog to digital Lucid AD9624converters, 
which are designed to work in a recording studio setting. They realized 
belatedly that the converter units themselves were not useful without 
a digitization station, which would cost several thousand dollars. The 
Center made the all too common mistake of purchasing equipment 
without a clear plan how the individual hardware or software will 
interface with existing equipment. Fortunately they were able to later 
purchase a digitization system that uses one of the Lucid converters. 

In order to ensure that in the future the Center made sound 
investments in technology and established a digitization workflow 
appropriate to their needs, the Director sought advice from oral 
history expert Doug Boyd at the University of Kentucky. Dr. Boyd 
visited LSU in March 2007 to evaluate the Center and conduct an 
introductory digital audio workshop for the Libraries staff. He 
generated a seven-page report with recommendations for equipment, 
collection development, and staffing. 

Recommended Analog to digital work station equipment 
and software 

1. Lucid AD9624 A/D Converter 

2. RME Hammerfall DSP 9632 PCI Audio interface 

3. 2 Yamaha HS50M 5" Active Monitor 



74 Digitization and Access of Louisiana Oral Histories 



4. 1 Tascam 202MKIII Dual Recorder Cassette Deck 

5. 4 BP20 20' TRS - TRS Cable 

6. 8 DKQR10 10' Dual RCA - TS Cable 

7. 1 Furman PL8II 15 Amp Power Conductor w/Light 
8. 1 DT770pro Closed Studio Mon Headphone 
9. 1 Presonus Cent. Station Audio Control Center 
10. 1 Plextor PX-716UF External CD-R/DVD+-RW 

11. Sony Sound Forge 8.0 Audio Editing Software 

12. Sony Noise Reduction 2.0 Noise Reduction Plug-In 

The equipment recommended in the report was purchased with 
endowment funds nearly a year after Boyd's initial visit. Boyd 
returned to the Center to help set up the equipment and train a newly 
hired staff member. 

Although not all institutions have the funds to hire a consultant, 
this less than $2,000 expenditure has proved money well spent for the 
Center. Without the vetting of the digitization program, the listed 
recommendations for equipment, and Boyd's 
encouragement to pursue positioning the Center as a leader in 
digitization efforts in the state and the profession, the Libraries 
administration might not have acted so quickly to support the 
endeavor. The administration approved reallocating funds to hire a 
full time staff member for the digitization and in less than two years, 
the Center has been able to digitize over 700 hours of interviews with 
their single digitization station. 

With the addition of a dedicated digitization station and full time 
staff member to manage the process, the Center was ready to begin 
digitizing in earnest. It was immediately apparent, though, that server 
space and file redundancy would be an issue. The average file size of 
one hour of digitized uncompressed audio from analog tape is around 
1.5 Gigabytes (GB). The Center only had access to a relatively small 
74GB drive when digitization began. 

Working with the Digital Services Librarian and the Libraries 
Systems Administrator, the Center temporarily located all digital 
audio files to a 5TB networked server that primarily serves as storage 
for TIFF images. In late 2009, a regional corporation donated used 



Digitization and Access of Louisiana Oral Histories 75 



storage equipment to the Libraries. The Libraries' Systems 
Administrator was able to configure four 2TB Raid 5 storage arrays, 
totaling approximately 8TB, for the Center's long term storage. This 
unexpected gift enabled the Center to continue digitization efforts, 
although they will still have to be selective. 

The Center exists not just to archive, but to conduct research- 
based oral history interviews and to educate the community about 
conducting interviews. To fulfill this mission, the Center keeps a stock 
of digital audio field recorders to loan for oral history projects. As 
noted earlier, this equipment is purchased with endowment funds. 
The Center currently has four Edirol R-09 recorders, two Marantz 
CDR 310 recorders, and five Zoom H2 Handy recorders for loan. 
Center staff uses a Marantz PMD 661 for interviews. 

The Edirol R-09 and Marantz CDR 310 are portable CD recorders 
and the Zoom H2 Handy records employ flash memory. Individuals 
borrowing the equipment are trained and instructed on its use. 
Digitally recorded interviews are brought to the Center either on CD 
or on secure digital (SD) flash memory cards. Interviews are saved to 
the Center's server and eventually processed. 

Digitization Workflow 

The digitization process is handled by one staff member, although he 
has recently trained a student worker to help run the digitization 
station. The staff member samples the audio to determine the optimal 
hardware and software settings and reformats the analog tape to a 
lossless uncompressed digital master WAV file. This master file is 
captured at a bit depth of 24 and a sample rate of 96 kHz in stereo. 

The master WAV file is stored on a networked server, which is 
routinely backed up to a tape drive. This "master file" is not altered 
after the initial digitization process. Whenever possible, barring any 
time or funding constraints, a copy of every collection is also stored on 
an external hard drive as well as burned onto a gold archival CD. 

The staff member then creates an optimized file from the master 
WAV file. Using Sound Forge software, he improves the signal 
strength and removes distortion from the audio. The optimized file is 



76 Digitization and Access of Louisiana Oral Histories 



saved as a WAV file to a different location on the server. He then 
generates a compressed MP3 file from the optimized file. This MP3 
file is the use copy, and it is also saved to the server. 

Unprocessed collections are digitized prior to processing to 
facilitate time stamping of the transcriptions. The Center uses Express 
Scribe Transcription Playback Software (http://www.nch.com.au/ 
scribe/) and adds time stamps to the transcriptions based on the 
actual run time. Old transcriptions will be re-audited and time stamps 
added because the tape time stamps are arbitrary, often reset every 
time the tape player is used. 

Metadata for the entire collection is kept in a Microsoft Access 
Database. All oral histories entering the center are processed based on 
a 13 page processing checklist. The processing checklist steps include 
1) Accession 2) Transcribe 3) Audit 4) Send to Interviewee 5) Edit. 
This process is time-tested and thorough. The majority of the oral 
history collection is cataloged according to AACR2 standards in 
MARC format in the LSU Libraries online catalog (i.e., OPAC). The 
Dublin Core metadata in the digital collections is often copied directly 
from these catalog records. 

Implementation and Access 

The Center does not currently have a formal collection development 
policy to determine which oral histories are digitized The interviews 
that have been digitized thus far were identified as "high risk" on 
unstable medium or they were considered to be of particular interest 
to researchers and the public. Materials are also digitized "on 
demand" for patrons for a fee. 

Tapes that were created prior to the Center opening in 1991 and 
later donated were assessed for deterioration and digitized as a means 
of preservation. For example, the 60 interviews in the Americans in 
Vietnam series, recorded between 1974-1977, were identified as at risk 
and were prioritized for digitization. Because of the content of the 
interviews, however, the digitized audio will not be offered via the 
LDL. In this situation, preservation of the materials outweighed the 
need to provide access. 



Digitization and Access of Louisiana Oral Histories 77 



Particular interviews and/or series of interviews, such as the 
Hurricane Betsy Series or the McKinley High School Series, were 
digitized because of their potential value to researchers and the 
general public. These collections will be uploaded to the LDL as soon 
as they are fully processed. Patron requested interviews that were 
digitized on demand for a fee are also candidates for the LDL. 

During the past two years the Center staff and Digital Services 
Librarian have discussed workflows for uploading audio to the LDL. 
They consulted collections mounted by the University of Louisville 
(http://digital.library.louisville.edu/), Ball State University 
(http://libx.bsu.edu/), University of Nevada, Las Vegas 
(http://digital.library.unlv.edu/), and the University of California, San 
Diego(http://ceo.ucsd.edu/index.html) to facilitate decision making 
about the organization and display of online oral history materials. 

The LSU Libraries serves all digital library materials via the 
Louisiana Digital Library, which was developed at the start of this 
decade by LSU Libraries and the LOUIS Library Consortium. LOUIS 
staff maintains the LDL for the nineteen participating institutions, 
including historical societies, libraries and museums. Individual 
institutions add content to the LDL and all materials are available for 
public use. The digital library is powered by CONTENTdm software 
and hosted by OCLC. LOUIS staff assists LDL institutions with 
customization of the software. LSU Libraries Special Collections 
maintains over 35 collections in the LDL. 

Adding audio collections to the LDL has been a slow process that 
seems to move in fits and starts. Center staff and the Digital Services 
Librarian have held many meetings and exchanged numerous emails 
about serving digitized oral histories online. Debate about the topic 
centered around how the interviews would be organized and 
displayed. Many interviews, especially the life narratives, are topically 
related even though they are in different series. For example, 
university history overlaps with civil rights history in several 
interviews. Organizing the interviews both topically and by series can 
be achieved by using CONTENTdm custom queries to unite items 
from different digital collections, although this method does require 



78 Digitization and Access of Louisiana Oral Histories 



staff to re-create the collection custom queries and topics or series are 
added. 

The CONTENTdm software seems more suited for its original 
purpose to serve digital images, and the default treatment of audio 
files is rather clunky. Audio does not play automatically, but instead 
the text "Access this item" appears at the top of the screen and 
metadata for the item below it. This presentation of the audio is 
somewhat confusing, because it is not even immediately clear that it is 
an audio file. Some institutions using CONTENTdm have devised 
workarounds that make serving audio in the software more usable. 

In order to better group interviews together with the 
transcriptions and other related content, the Digital Services Librarian 
began uploading files as "compound objects" or multi-part files in 
CONTENTdm. Figure 2 illustrates this with the different files, 
abstract, transcription, and audio, hyperlinked in the left column. This 
display is not ideal since the metadata for the interview is on a 
separate screen and the "Access this item" text is still present. In 
addition to the cumbersome nature of the audio display, patrons 
wishing to listen to it are forced to download the often very large file 
to their computer. The Director felt strongly that other options not 
requiring the patron to download the audio be explored. Copyright 
would be difficult to manage if the audio was copied to different 
computers. 

After reading about Ball State University development of a user- 
friendly embedded Windows Media Player above the PDF file within 
CONTENTdm (Hurford & Read, 2008), the Digital Services Librarian 
contacted LOUIS about implementing this method. LOUIS staff 
worked with the LSU Information Technology Services (ITS) 
department to obtain access to a streaming server from which the 
audio could be served. MP3 files are uploaded to the server via FTP 
software and the file path is linked to the item in CONTENTdm in the 
metadata field "Stream File". 

The embedded player facilitates ease of use by providing the 
searchable PDF transcription to the patron as they listen to the audio. 
It does not require listeners to download the audio, thus it better 



Digitization and Access of Louisiana Oral Histories 79 



protects the copyright of the files. Information about copyright is 
included in the metadata for each item and future transcriptions may 
be watermarked with a copyright statement. 

To organize the oral history collections in the LDL, the Digital 
Services Librarian used the "collection of collections" model that 
CONTENTdm employs to organize user collections on their website 
(http:/ /www.oclc.org/ contentdm/ collections/default.htm). The 
individual series or collections are cataloged as a whole in the overall 
Center LDL collection. The series are represented by an image and 
selecting that image displays metadata taking the patron to the 
interviews. CONTENTdm software allows the creation of custom 
queries that will link the different collections and enable patrons to 
search across them. The individual series can be added to and the 
interviews and other materials in the collections will remain together, 
searchable alphabetically by title. 

Problems and Some Solutions 

Every digitization endeavor has its problems, but it is the individual 
institution's staffing, resources, and prior experiences that dictate the 
solutions. The Center, although small, is supported by a large 
university library. Digitization is a luxury that can be afforded because 
the Center has endowment money to purchase equipment and to 
provide staff with continuing education in the field. The time it takes 
to digitize resources is not a major factor in the continuation of 
digitization either because digitization is accepted as a part of the 
overall processing workflow. Digitization at the Center will be funded 
indefinitely and a full time employee will be dedicated to the effort if 
at all possible. 

The Center is now two years into their programmatic digitization 
effort. At this point the digitization workflow has been well established 
and interviews from a few collections have been uploaded to the LDL. 
This section of the paper details problems encountered, such as legacy 
digitized collections, prioritizing digitization efforts, storage solutions, 
staffing, and digital access and display via CONTENTdm software, 
and how the Center staff and the Digital Services Librarian resolved or 



80 Digitization and Access of Louisiana Oral Histories 

did not solve them. Many problems could have been mitigated with 
more long-term planning, but the degree to which digitization efforts 
are currently supported and the ramifications of beginning a 
digitization program were not known at the start of these efforts. 

The Center holds some legacy digital collections that do not meet 
the current standards for digitization. Prior to acquiring the 
digitization station and hiring an audio technician, Center staff did 
some preliminary digitization of analog tape using an external cassette 
tape deck connected to a computer. The sound was collected using a 
low end sound card to ram (Real Media Player) format in a process 
like the one that Washington State University Libraries used for their 
African-American Oral History collections (Bond, 2004). These early 
recordings were deemed important enough to place in the queue to be 
re-digitized according to the Center's current standards. For practical 
purposes, an institution may choose to keep legacy digitized items 
even if they do not meet current standards because the cost to re- 
digitize is high. For the Center the lessons learned with early 
experiments in digitization were important in shaping the future 
decisions to allocate more funds and staff to the digitization efforts in 
order to produce better quality sound. 

The Center's at risk materials were digitized first, however, some 
of these materials are not good candidates for online access. The files 
will need to be stored long term, but because of restrictions they will 
be largely inaccessible. This falls within the mission of the Center, 
which includes collecting in addition to providing access to oral 
histories. Some audio files do not have completed transcriptions, 
rendered them unacceptable for immediate uploading to the LDL. The 
interview editing process is very time consuming and there is little 
immediate results (Bond and Walpole, 2006). Digitization priorities 
may differ depending on the institutional mission. If the mission is to 
provide access and preservation is secondary, then more popular or 
relevant collections should be digitized first. Institutions not 
supported by a parent institution, such as the Center is by LSU 
Libraries Special Collections, may not have the luxury to digitize 
collections just to archive them. 



Digitization and Access of Louisiana Oral Histories 81 



Another ongoing issue is long term storage solutions for the 
digitized files. The Center hoped to have files saved in at least three 
different places, a dedicated server in the main library, CD, and offsite 
storage. Some files are saved to an external hard drive in addition to 
the networked server, and born digital audio is saved to Gold CD. 
Ideally a copy of each master WAV file would be stored in offsite 
storage in a similar set up to the University of Kentucky (Weig, Terry 
& Lybarger, 2007), but this has not been implemented. The Libraries' 
server on which all audio files are saved is backed up incrementally to 
magnetic tape every night. Full backups take 40-120 hours because of 
the amount of data contained on the servers, so they are conducted 
once monthly. It is a secure system, but there is always a chance for 
failure. Future plans call for the Center to assess file storage and 
redundancy options. 

The document "Sound Directions: Best practices for digital audio 
preservation" provides recommendations for long term preservation 
storage (Casey & Gordon, 2007), however many recommendations 
may not be feasible for small centers. The authors emphasize that file 
redundancy which is neither labor-intensive nor costly in media (e.g., 
CD or flash memory), should always be implemented. The majority of 
institutions will likely not have multiple terabyte servers and staff to 
keep them running, but files can at the very least be backed up to a 
more affordable storage medium such as portable hard drive or CD. 
Any storage medium can fail, however, so careful attention to this 
matter is imperative if an institution is interested in long term storage 
of files. 

An issue that may require further review and assessment is the 
current standard of capturing audio at the higher sample rate of 24 bit 
96 kHz. As server space fills and the Center and Libraries' budgets 
decrease, however, this standard may be reduced. Capturing audio at 
16 bit 44.1 kHz reduces the file size by nearly half, and according to 
some experts it does not substantially decrease the quality of the WAV 
file (Weig, Terry & Lybarger, 2007). If the server is filled the Center 
may elect to save the derivative optimized WAV file to CD rather than 
the server. File optimization is time consuming, often taking the 



82 Digitization and Access of Louisiana Oral Histories 

length of the recording to complete, so deleting these files is not an 
option. 

Before embarking on a digitization project, an institution should 
estimate the number of files that will be created and storage space 
needed. An institution may choose to capture audio at a lower and still 
acceptable rate to expedite the digitization process and conserve 
storage space. The institution should conduct an assessment of 
whether file optimization and multiple WAV files are needed before 
creating additional files that must be saved over the long term. Any 
derivative files can be recreated, so they should always be deleted or 
copied to more affordable storage media if server space is at a 
premium. 

Another issue related to the audio capture standards is the 
Center's lack of written standards and best practices. Workflow 
principles and digitization methods are generally adhered to, but there 
is no guide or manual, just institutional knowledge. The workflow is 
based on recommendations by oral history expert Doug Boyd, who 
served as an advisor to the Center and also wrote the tutorials and 
information found on the Oral History Association website 
(http://www.oralhistory.org/technology/). The Center should apply 
the same level of detail and documentation to digitization workflow as 
they have for the processing workflow. 

There are only two staff members who work with the Center's 
digital files, which could pose potential problems if either leave and 
has often caused bottlenecks in the workflow. At the Center all 
digitization is handled by one staff member with some student 
support. Other Center staff members do not have time to perform 
these duties, so little cross-training has been done. This is a risk 
because if the staff person leaves it will be difficult to continue 
digitization efforts. In the same vein, only the Digital Services 
Librarian currently uploads items to the LDL. This duty is usually 
shared by graduate assistants, but financial constraints have 
prevented hiring any additional help. Digitized files often do not get 
uploaded quickly because they are placed in a queue with all Special 
Collections digital materials. Cross-training between the digital 



Digitization and Access of Louisiana Oral Histories 83 



technician and Librarian is an option that should perhaps be explored. 
At the very least the two individuals, who are separated geographically 
across campus should establish better communication and more 
effective workflows. Information about which collections are ready to 
be uploaded to the LDL is sent ad hoc via email and there is no current 
mechanism for tracking the LDL files via the Accession database. 
Institutions should establish a clear workflow and assign 
responsibility for different aspects of the digitization process early on 
in a project. This will alleviate any potential miscommunication or turf 
war situation. 

Before purchasing equipment and hiring staff to digitize audio, an 
institution should assess the environment where they will be located. 
At the Center the digitization station is equipped with the right 
hardware and software, but its location is less than ideal. The Center is 
located in an 80 year old house that is poorly insulated. The room in 
which digitization takes place is in the center of the house next to the 
building air handlers. The sound technician must use headphones 
while optimizing audio. If the Center is relocated much thought will be 
put into the location of the digitization station. In addition, Dr. Boyd 
recommended the Center purchase two digitization stations. When 
funding is available, the Center will explore this option. 

A very important aspect of digitization efforts is providing access. 
The Center works with the Digital Services Librarian to upload items 
to the LDL, which uses CONTENTdm software. The software is less 
than perfect in its treatment of audio files, and efforts to retrofit the 
software to better serve audio are time consuming dependent on 
LOUIS staff expertise. LOUIS controls server access so software 
customization must go through them. The Center benefits from being 
a part of this consortium environment where an infrastructure is in 
place and support is offered at all times, but there are some 
constraints that this relationship brings. Small or not well-funded 
institutions interested in mounting collections online may be better 
served entering into a partnership with a larger institution or 
consortium. 



84 Digitization and Access of Louisiana Oral Histories 

An issue specific to the retrofitting of the software potentially 
affects patron access and sustainability. The embedded audio player 
that LOUIS retrofit for audio display does not display a time stamp so 
patrons cannot skip to a specific section of the interview. The audio 
player works well in the most current version of CONTENTdm, but 
the software is scheduled to be upgraded soon. Changes may affect the 
workflow and change the player functionality. The Center will rely on 
LOUIS consortium staff to recreate the embedded player in the 
upgraded software. Some institutions may not be able to expend a 
great deal of staff time continually addressing the interface when the 
software is upgraded, so this should be considered when addressing 
the sustainable access points. 

In many ways the process for adding audio collections to the LDL 
has just begun. In 2008 all processed oral history collections which 
had been on a cataloging backlog were cataloged in the Libraries 
OPAC and WorldCat, which facilitates the metadata creation of 
records in the LDL. Changes in the CONTENTdm software in the past 
few years have made it more customizable. In 2009 the Center staff 
began producing audio and video podcasts with images and sound 
from the collections. The podcasts and information about hem are 
available on the Center's blog (http://oralhistory.blogs.lib.lsu.edu/). 
In order to maximize the amount of digitized materials that are 
available online, key players should outline a digital access plan 
wherein all materials that are currently ready for public display are 
listed and other materials are prioritized. 

Conclusion 

The T. Harry Williams Center for Oral History began a digitization 
program a little more than two years ago. Since then the Center has 
acquired digitization hardware and software, hired a full time staff 
member to perform digitization duties, and mounted several 
collections to the Louisiana Digital Library. By all accounts, the 
Center's efforts have been successful, although they hope to develop 
more sound workflows for digital access to enable them to add 
additional interviews to the online collections in the future. 



Digitization and Access of Louisiana Oral Histories 85 



Institutions wishing to emulate the Center should consult experts in 
person or through the literature, follow industry standards set forth by 
the Oral History Association (http://www.oralhistory.org), and, 
formulate plans based on best practices such as the CDP Digital Audio 
Working Group Digital Audio Best Practices 
(http://www.bcr.org/dps/cdp/best/digital-audio-bp.pdf). It is 
essential to plan ahead for storage space needs, keeping in mind that 
what one thinks you'll need is probably less than the reality. 

References 

Bond, T. J. & Walpole, M. (2006). Streaming audio with synchronized 
transcripts utilizing SMI., Library Hi Tech 24, 452-462. 

Bond, T. J. (2004). Streaming audio from African-American oral 
history collections. OCLC Systems & Services, 20, 15-23. 

Casey, M. & Gordon, B. (2007). Sound directions: best practices for 
audio preservation. Retrieved from: http://www.dlib.indiana.edu/ 
projects/sounddirections/papersPresent/sd_bp_07.pdf 

Hurford, A. A. & Read, M. L. (2008). Bringing the voices of 

communities together: the Middletown digital oral history project. 
Indiana Libraries. 27, 26-29. 

Integration and the black experience. (2003). Retrieved December 14, 
2009 from http://www.louisianadigitallibrary.0rg/cdm4/ 
browse.php? CISOROOT=/IBE 

T. Harry Williams Center for Oral History Exhibits and Presentations. 
(2009). Retrieved December 14, 2009 from 
http : / / www. lib .lsu.edu/special/ williams / ep .html 

The Baton Rouge Bus Boycott 0/1953. A recaptured past (2004). 
Retrieved December 14, 2009 from http://www.lib.lsu.edu/ 
special / exhibits /boycott / index.html 

Weig, E., Terry, K. & Lybarger, K (2007). Large scale digitization of 
oral history: A case study. D-Lib Magazine 13. Retrieved from: 
http:/ /www. dlib.org/ dlib/ may07/weig/ 05weig.html 



Digitizing a Newspaper Clippings 
Collection: a Case Study and Framework 
for Small-Scale Digital Projects 

Maureen M. Knapp (John P. Ische Library, New Orleans) 



Abstract 

How does a small specialty library establish, develop and maintain in- 
house digital collections? What are the considerations, challenges, and 
benefits they experience? This chapter describes one library's 
experience in turning an aging and inaccessible collection of 
newspaper clippings into a preserved and searchable online collection, 
which in turn laid a basis for other digital projects. This chapter also 
discusses considerations, challenges and opportunities observed 
during their first foray into creating a digital collection. 

Keywords: Clippings, Digital libraries, Digital preservation, Digital 
projects, Digitization, Electronic preservation, Newspaper clippings 
file, Newspaper clippings, Press clippings. 



Background 

The John P Ische library is a mid-sized, urban, academic health 
sciences library serving six schools of health professions at the LSU 
Health Sciences Center (LSUHSC) in New Orleans, Louisiana. 
Established in 1931, the library has collected newspaper clippings 
related to the history and accomplishments of the health sciences 
institution since its inception, and even today monitors the local 
papers for pertinent news items. The "newspaper clippings file," as it 



86 



Digitizing a Newspaper Clippings Collection 87 



came to be called, is an astounding 70 year snapshot of the 
development of the health sciences in Louisiana. Over 6,000 clippings 
trace development of LSUHSC through the twentieth century, 
including such topics as: the people, places and events associated with 
the LSU School of Medicine, the growth of health infrastructure in 
Southeast Louisiana and New Orleans, and the development of 20th 
century health sciences education in Louisiana. 

Digital Collection Origins 

In 2002, access and preservation concerns with some of the earliest 
newspaper clippings encouraged the library to investigate digitization 
as a possible solution. Access points to the collection were limited. The 
only online access consisted of a locally-created subject database 
containing basic citations to newspaper articles from 1985 to present. 
Users had to search the local database by faculty name or department, 
and then locate the physical newspaper clippings in filing cabinets by 
call number. The remaining fifty-odd (1933-1984) years of the 
collection was indexed in a card catalog, stored in the library's back 
offices and only accessible to library staff. 

Numerous problems plagued the physical collection. The 
newspaper clippings had been stored in filing cabinets as they were 
collected, which allowed the typing paper to curl heavily over the 
course of many years. The newsprint itself showed signs of age: rust 
marks appeared where staples and paperclips had once connected 
pages, and gaps in the collection were apparent. 

A lack of funding and staffing was another concern. Any efforts 
towards creating a digital collection would have to be inexpensive and 
make use of staff and resources the library already possessed. 

However, to truly understand to physical condition of the 
newspaper clippings file, and the challenges that would arise once 
digitization began, one must understand the collection process of 
gathering the original newspaper clippings. While no documentation 
exists, the library postulates that even back to the 1930s, a library 
member would skim the daily local papers from around Southeast 
Louisiana for any mention of LSU School of Medicine, and its faculty, 



88 Digitizing a Newspaper Clippings Collection 

staff or students. Once an article was discovered, it was cut out of the 
paper, dated, and the name of the paper was noted. The articles were 
glued to standard 8 V2 by 11 inch typing paper, usually several to a 
page, somewhat in order by date, and the paper was assigned a 
numerical call number in the order they were received. Later someone 
would read the articles, underline named entities pertaining to LSU, 
and assign a subject heading, which was recorded in a small local card 
catalog. Finally, the pages of clippings were organized into manila 
folders by year and placed into filing cabinets until further needed. 
This entire process continued for 50 years. 

So basically, the library had a unique local news collection, 
spanning the majority of the 20th century, collected and stored under 
questionable archival methods, with limited access to documents 
before 1985. In order to increase availability and use of the clippings, 
the library wrote a grant proposal for a small-scale digitization project 
to scan the newspaper clippings from 1933-1953, streamline 
cataloguing, and offer public access to the resource online. The grant 
proposed using Greenstone digital library software, an open source 
"suite of software for building and distributing digital library 
collections" (Greenstone digital library software, 2007), to provide 
access to the digitized newspaper clippings. 

Stops and Starts 

Though the grant proposal was rejected, the grant writing process did 
provide a catalyst for action within the library. The small grant 
requested $3,000 to purchase a flat-bed scanner, computer and 
optical character recognition software. Library administration was 
impressed enough with the grant's digitization plan that they provided 
funding for a scanner, software and travel to a continuing education 
class on digital projects in 2003. A library staff member began 
scanning the clippings. However, the library quickly ran into 
problems. The Greenstone software would not work properly on their 
secure intranet, and the library lacked a staff member with enough 
computer programming experience to install and troubleshoot the 
software properly. In addition, the image quality of the scanned 



Digitizing a Newspaper Clippings Collection 89 

newspaper clippings was poor, which was attributed to a faulty 
scanner that did not produce dark enough images. Finally, copyright 
concerns made library administration hesitant to post the collection 
online to the general public. 

By the time Hurricane Katrina struck New Orleans in August 
2005, access, software and image quality issues had put the library's 
newspaper clippings digitization project on hold. The library's 
collection was undamaged from this natural disaster. However, it was 
moved to remote storage for over half a year and the entire library 
staff was displaced. 

During the ensuing hiatus, library staff took several continuing 
education classes on digitization. "Digitization Fundamentals," a 
course offered by the Illinois Digitization Institute at the University of 
Illinois Urbana-Champaign (University of Illinois Library, 2009), was 
exceptionally useful, as it provided training in digital projects 
management, standards and organization, as well as an introduction 
to Photoshop software. 

In 2007, an opportunity opened for the library to join the 
Louisiana Digital library, the state digital library consortium provided 
through LOUIS: The Louisiana library network (LOUIS: The 
Louisiana library network, 2009). The library was able to obtain 
access to OCLC's CONTENTdm platform, which was previously too 
expensive, as well as the technical infrastructure and support needed 
to store and access digital assets. 

Consortial membership for digital library services addressed 
many of the problems faced by the library developing an in-house 
digital collection. The documentation on the technical and operational 
requirements for participation in the LOUISiana Digital library 
proved critical. The consortium's style manual for scanning and 
cataloguing provided guidelines for selecting collections to digitize, 
scanning practices, post-scanning image manipulation, project 
workflows, metadata standards, and quality control. Another practical 
advantage to consortial membership was LOUIS staff support, which 
provided advice on imaging standards, basic training on the 



90 Digitizing a Newspaper Clippings Collection 



CONTENTdm software, and a shoulder to cry on when things went 
awry. 

The library began their second try at developing a digital version 
of the newspaper clipping file in January 2008. As of December 2009, 
the library has not only met their original goal of digitizing and 
indexing over 1600 items in the collection from 1933-1953 (LSUHSC 
New Orleans library, 2009), but also created several other collections. 

Work Flow, Image Manipulation and Standards 

The format and organization of the newspaper clippings collection 
created a challenge in regards to digital manipulation and workflow. 
In order to achieve indexing of items on an individual level, some 
information that was included only once on a sheet of several 
newspaper clippings (for example, the name of the newspaper, the 
date, and most commonly, the clipping's call number) would have to 
be added to each individual item. Thus, several steps beyond simple 
scanning and image processing were included in the workflow. 

Here are the workflow and standards for creating digital versions 
of the Newspaper Clipping File: 

1. Following consortium standards for creating digital images for the 
Louisiana Digital library, the full-page newspaper clipping is 
scanned on an HP Scanjet 8390 flatbed scanner to create an 
archival black and white image at 300 dpi, 8-bit grayscale and 
saved as an uncompressed TIFF file on the library server. This 
creates an archival master version of the original digital image. 

2. Using Photoshop, a copy of the archival master version is opened 
and saved according to file naming conventions for the digital 
library set forth by the consortium. This creates a duplicate of the 
archival master that can be manipulated to isolate an individual 
clipping. This file is the image that will eventually be loaded into 
the digital collection. 

3. The duplicate is cropped to isolate a single newspaper clipping. 
Pages that have only one clipping on them are also manipulated 
and cropped to minimize file size. 



Digitizing a Newspaper Clippings Collection 91 



4. If not visible, the call number, date and newspaper name from the 
original scan are copied, cut and pasted to the now isolated 
clipping. 

5. Post capture processing is applied. The item is processed for 
alignment and an unsharp mask filter is applied to correct blurring 
that might have occurred during the scan process. In addition, the 
image's histogram is viewed to adjust color intensity. 

6. The individual, processed image of the individual newspaper 
clipping is saved to the server. 

7. For pages with more than one newspaper clipping, this process is 
repeated until all clippings have been isolated. 

8. After digital manipulation, the TIFF of the clipping is loaded into 
the CONTENTdm Project Client. Cursory metadata is entered by a 
library staff member. The file name, size and location are recorded 
in a Scanning Log to track progress. 

9. The librarian performs Optical Character Recognition (OCR) on 
the clipping to create an excerpted text field and assigns subject 
headings. OCR produces an abstract of the first 50 words of the 
article, which is keyword searchable in the digital library. This 
takes a bit of time, but it is a good way to review the article and 
assign the proper subject heading. After a final quality check, the 
item is approved and uploaded to the digital library. Upon upload, 
CONTENTdm converts the full resolution TIFF file to JPEG, 
which is what end-users access when viewing the collection online. 

10. CONTENTdm also offers an Archival File Manager, which 
automatically archives collections in a location specified on our 
library server as they are uploaded to the online collection. Once a 
volume is full, it is burnt to an archival quality CD recordable disc, 
as well as saved on the server. 

Cataloging and Metadata 

The LOUIS consortium requires collections in the Louisiana Digital 
library to use the Dublin Core 15 metadata element set (Dublin Core 
Metadata Initiative, 2008), in addition to non-Dublin core structural 



92 Digitizing a Newspaper Clippings Collection 



and administrative metadata. CONTENTdm allows up to 125 fields 
per collection. The library decided to add 3 more metadata fields to 
the newspaper clippings collection: Call number (to locate the item in 
the physical files), Full Text (for excerpted text) and Contact 
Information (so users can contact the library). The following lists the 
metadata fields used in the newspaper clipping collection. 



Field Name (in 


Type of 


Metadata Content 


Added by 


Title 


DC 


Title of newspaper clipping 


LS 


Contact Information 


A 


Contact information for library 


T 


Creator 


DC 


Author of clipping 


L 


Contributors 


DC 


Contributor to clipping (rarely used) 


L 


Subject 


DC 


Institutional controlled vocabulary, 
MeSH 


L 


Call Number 


D 


Call number for the original clipping 


LS 


Description 


DC 


"Newspaper clipping" 


T 
1 


Notes 


D 


More descriptive information about 
content of original clipping, if needed 


L 


Publisher 


DC 


Newspaper title 


L 


Date 


DO 


Date of publication 


L 


Type 




1 exi 


T 
1 


Format 


DC 


"TIFF" 


T 


Identifier 


DC 


Mandatory field directs users to 
identifier URL 


T 


Source 


DC 


Library name and homepage URL 


T 


Language 


DC 


"En." 


T 


Relation 


DC 


URL to homepage of Newspaper 
Clippings Collection 


T 


Coverage - Spatial 


DC 


"New Orleans (La.)" 


T 



Digitizing a Newspaper Clippings Collection 93 



Field Name (in 


Type of 


Metadata Content 


Added by 










Coverage - 
Temporal 


DC 


Year of publication 


L 


Rights 


DC 


Copyright information 


T 


Cataloger 


D 


Initials of librarian 


L 


Cataloged Date 


D 


Date of cataloging 


L 


Object File Name 


D 


File name of item 


LS 


Image Resolution 
(Archival) 


A 


Dots-per-inch of scanned TIFF i.e.: 
'300dpi" 


T 


Image Bit-Depth 
(Archival) 


A 


"8-bit" 


T 


Color Mode 
(Archival) 


A 


Grayscale 


T 


Extent (Archival) 


A 


Pixel dimensions of image 

/l A /l A /l A /l A / 1 II II II 1 \ 

(WWWW:HHHH) 


LS 


Image Manipulation 
(Archival) 


A 


"Crop, alignment, unsharp mask, 
histogram" 


T 


File Size (Archival) 


A 


Size of TIFF image in KB 


LS 


Hardware / Software 
(Archival) 


A 


HP Scanjet 8390, Photoshop, ABBYY 
FineReader" 


T 


Digitized By 


A 


Initials of library staff member 


LS 


Digitized Date 


A 


Date of digitization 


LS 


Full Text 


D 


Abstracted content from OCR 


L 



List of metadata elements used in cataloging items. Meaning of symbols: A is 
administrative; D is descriptive; DC is Dublin Core 15; LS is added by Library Staff, L 
is added by Librarian, and T is added by Template. 



Many of these fields are inserted automatically via a template in 
CONTENTdm. The remaining fields are divided among project 
members. The most tedious data entry was entering the Extent and 
File Size fields for each item. Each clipping's dimension and size is 



94 Digitizing a Newspaper Clippings Collection 



different, so library staff tends to write these down on a notepad as 
they scan images for entry, then record them in CONTENTdm and the 
scanning log later. 

Another feature of Content DM is the ability to build a customized 
controlled vocabulary for the Subject field. This worked to the library's 
advantage, as the newspaper clipping file possessed a card catalog of 
subjects. The library uses the newspaper clippings card catalog as a 
basis to build an institutional controlled vocabulary in the digital 
library. The card catalog also serves as a reference point to verify 
names and spellings of affiliated persons. This institutional controlled 
vocabulary can be shared across digital collections, which is an 
advantage for future projects related to our institution. 

The library soon recognized that other subjects would be 
necessary to adequately describe the digitized newspaper clippings. 
Original cataloging varied so much over the years that clippings might 
only include the name of the person or entity mentioned in the article. 
The library wanted to add more descriptors, so that articles describing 
conferences, publications, research grants or other common topics 
were easier to locate. When applicable, the library consults the 
National Library of Medicine's list of Medical subject headings 
(MeSH)(U.S. National Library of Medicine, 2009) for appropriate 
descriptors in the Subject field. For example, the MeSH term 
"Congresses as Topic" is used when a clipping discusses conferences, 
or the MeSH term "Publications" when a clipping mentions a new 
book or journal article published by one of the institution's faculty. 
Sometimes, MeSH is not useful, especially when discussing local 
events such as campus expansion or departmental news. In these 
cases, a subject heading is created and assigned by the librarian. 
Clippings in the digital collection can be browsed by year, subject, 
creator or title. Browsing by date is an interesting way to view the 
development of institutional history. To further open the collection, 
keyword searching is enabled in the excerpted text field. 



Digitizing a Newspaper Clippings Collection 95 



Project Considerations 

Storage, standards, documentation, training and staffing were all 
considerations for this project. 

Storage was a huge concern. The deteriorating condition of older 
newspaper clippings made it evident that storing the physical 
newspaper clippings in filing cabinets was not conducive to 
preservation. To address the curling paper, books were used to weight 
down the paper for several weeks. This did not entirely fix the issue of 
curling paper, but it did help a little in preparing the clippings for a 
move to flat storage. After flattening, the files were transferred to acid- 
free archival folders and placed in clamshell archival storage boxes. 
Finally, the clamshell boxes of physical files were relocated to the 
library's humidity controlled Rare Books Room, in order to protect 
them from humidity and sunlight. 

Likewise, the library was heedful of digital storage and the "digital 
mortgage": how will the library address transfer of archival TIFF files 
to new formats as software and hardware change? Though the library 
has yet to encounter a change in image format standards, they did 
attempt to prepare for this inevitability by storing the collection of 
archival images in multiple locations, as well as on multiple formats. 
Having multiple copies also addresses the possibility that some files 
might eventually become corrupted. TIFF versions of the images are 
burnt to an archive quality, professional grade CD recordable discs, as 
well as copied to a location on the library server, which is maintained 
by our institution and backed up daily to tape at a remote location. 
This is in addition to the processed JPG file that is available to the 
public on the Louisiana Digital library. A TIFF of the raw scan of the 
original newspaper clipping is also retained on the library server. 

With multiple storage locations and a complicated workflow, 
documentation and staff training are also important concerns. The 
library's consortial membership provided a style manual for scanning, 
cataloging/metadata standards, and basic workflow suggestions. The 
library used this as a basis for creating a local workflow policy, which 
includes detailed directions on image scanning and manipulation as 
well as step by step directions on how to process the item in 



96 Digitizing a Newspaper Clippings Collection 

CONTENTdm. A scanning log is used to track size and progress of a 
collection. The scanning log is simply an Excel file which records the 
file name, file size, and date of digitization, as well as locations to 
which the file has been saved. 

Regarding training, the library realized it was critical that 
everyone involved with the project learn Photoshop. The LOUIS 
consortium takes a 'train the trainer' approach to CONTENTdm, so 
the librarian was responsible for training local staff on the software 
after initial training. 

This project is staffed with one librarian and two library staff 
members, who devote about 10 hours a week to this project. Library 
staff is requested to scan and process 60 clippings per week. 
Scheduling issues quickly became apparent for the librarian project 
manager, who has bibliographic instruction and reference desk duties 
in addition to overseeing digital projects. A supervisor suggested 
setting aside one day a week to solely devote to digital projects. Friday 
has since become "Digitization Day" and has worked well in keeping 
the load of items to be processed and approved by the librarian to a 
reasonable amount. 

Benefits and Challenges 

One of the first challenges was software sustainability. The free 
Greenstone digital library software did not work within the 
institutional intranet and required higher level technical skills than 
the library possessed. In addition, problems with the original project 
scanner resulted in poor quality images that had to be redone. 

Support from your institution from inception is critical. 
Administration has to be on board to provide funding and act as a 
liaison to other resources, for example, consulting with your 
institution's legal department about copyright questions. Support 
from information technology (IT) is also important. Getting our IT 
department to provide support for open source library software was a 
challenge that soon put the library's original plans to use Greenstone 
digital library software on hiatus. One of the benefits of membership 
in a state digital library consortium is that technical support is 



Digitizing a Newspaper Clippings Collection 97 



provided in an automated timely manner. In addition, the consortium 
has direct contacts with the software developers at CONTENTdm, so 
software concerns are quickly addressed. 

The newspaper clippings collection is unique in that it collects 
clippings from many regional news sources. All materials were 
published after 1923. Therefore, the work may be protected by 
copyright until 2018. Violation of copyright was a large concern, so the 
library decided to restrict access to the images within the newspaper 
clippings collection to the institutional IP address. In order to share 
the collection with a larger audience, the collection's metadata is 
searchable and viewable to anyone. This way, any user can find items 
in the newspaper clippings collection, and if they are not from the 
institution, the library works with them to get the information or 
clippings they need. 

Funding is a final challenge. Consortial membership to the digital 
library is about $2000 a year, while hardware and software ran about 
$1500 in startup costs. In addition, the library director donated a 21" 
screen won at a library conference raffle for use with the digital 
projects computer. Digital imaging is much easier with a larger screen. 
Grants and scholarships are another source of funding. A scholarship 
from a regional medical library group helped fund attendance at the 
first continuing education class on digital imaging and metadata for 
the librarian project manager. An recent Institute of Museum and 
Library Services "Connecting to Collections" Bookshelf grant 
(Institute of Museum and Library Services, 2009) allowed the library 
to obtain a set of conservation resources and books, which was 
previously non-existent. 

The library now has over 10 years of institutional history available 
online in a searchable database. Visibility and access to this collection 
has increased. Indexing though OCLC allows results to appear in 
Google. As a result, the library has received several inquiries about 
subjects indexed in the newspaper clipping file from the United States 
and Italy. The clippings file has also acted as a catalyst for change, 
inspiring library staff to organize the rare books room, research 
archival storage methods, and apply for grants. One of the benefits the 



98 Digitizing a Newspaper Clippings Collection 

library is proudest of is the mentoring opportunity this created. A 
library staff member who helped start this project recently completed 
their library degree and went on to become a Digital Initiatives 
librarian at another local library. 

The library has established a workflow and gained experience in 
digital imaging and management for future projects. Because of the 
success in creating the newspaper clippings collection, the LSUHSC 
School of Dentistry started a digital collection of historic photographs. 
In addition, the library worked with the LSUSHC Registrar's Office to 
digitize graduation program records, which are now available in a 
public, searchable collection. Finally, the library is in the planning 
stages of creating a digitized version of early volumes of the medical 
school student newspaper. The library also continues to add items to 
the newspaper clippings collection. 

As one can surmise, it has been a long 4 years to produce this 
digital collection, but once the library established workflow and 
standards it was much easier to begin other projects. Support from the 
state library consortium certainly expedited and streamlined the 
process, and the library recommends state or regional consortium 
membership to any smaller institution considering developing a 
digital project. For all the tedious data entry and malfunctioning 
software, the creation of an enduring, searchable and accessible 
source of institutional history made the entire project worthwhile. 

References 

DCMI Usage Board. (2008). DCMI type vocabulary. Retrieved 
December 9, 2009, from http://dublincore.org/documents/dcmi- 
type-vocabulary/ 

Dublin Core Metadata Initiative. (2008). Dublin core metadata 
element set, version 1.1. Retrieved December 9, 2009, from 
http:/ / dublincore.org/ documents/ dees/ 

Greenstone digital library software. (2007). Retrieved December 9, 
2009, from http://www.greenstone.org/ 



Digitizing a Newspaper Clippings Collection 99 



Institute of Museum and Library Services. (2009). Connecting to 
collections: A call to action. Retrieved December 9, 2009, from 
http://www.imls.gov/Collections/ 

LOUIS: The Louisiana library network. (2009). Retrieved December 
9, 2009, from http://apploo3.lsu.edu/ocsweb/louishome.nsf/ 

LSUHSC New Orleans Library. (2009). LSUHSCNew Orleans 

newspaper clippings collection homepage. Retrieved December 10, 
2009, from http://www.louisianadigitallibrary.0rg/cdm4/ 
index_LSUHSC_NCC.php?CISOROOT=/LSUHSC_NCC 

U.S. National Library of Medicine. (2009). Medical subject headings - 
homepage. Retrieved December 9, 2009, from 
http://www.nlm.nih.gov/mesh/meshhome.html 

University of Illinois Library. (2009). Digital services and 
development — training. Retrieved December 9, 2009, from 
http : //images .library, uiuc. edu/pr oj ects/ newpr oj .htm 



METRO Grant Success Story: Waterways 
of New York Project 



Claudia A. Perry and Thomas T. Surprenant 
(Queens College, CUNY.) 



Abstract 

The concept of experiential learning is particularly useful when 
students are required to create database entries as part of an ongoing, 
real-life, online experience. A METRO grant resulted in an 
opportunity to use students to create a CONTENTdm database which, 
with the continued software support from METRO, has continued and 
evolved until the present. This chapter describes the experience of 
both faculty and students. Sections include the background, technical 
issues and implications for teaching, project procedures and 
workflow, successes and lessons learned, challenges and next steps. 
Of particular interest is the use of out of copyright postcards and the 
metadata that has resulted from intensive student study and 
evaluation of the data contained on these cards. Those contemplating 
a digitization project of their own will be able to learn much about best 
practices, project planning, management and the 
advantages/disadvantages of the CONTENTdm software. 

Keyword: Best Practices, Canals, Case Studies, Cooperative 
Learning, Digitization, Digital Collection Management Software, 
Digital Collections, Digital Imaging, Experiential Learning, Library 
Education, Metadata, Postcards, Project Based Learning, Project 
Management, Project Planning, Quality Control, Standards, Student 
Developed Materials, Student Participation, Student Projects, 
Waterways. 



100 



Waterways of New York Project 101 



Introduction 

For many of us, hands-on learning is the best way to integrate an 
understanding of principles and best practices with a practical grasp 
of the actual challenges and learning opportunities of a project. This 
is particularly true for library school graduate students seeking to 
expand their theoretical, technical and management skills. As 
digitization is increasingly seen as a worthy endeavor for even the 
smallest institutions, it is worth considering the range of approaches 
available for gaining needed expertise, especially at the novice level. 
Examining the long-term development of an integrated, semester- 
long, course-based approach to digitization may be of value for those 
seeking an inexpensive approach for the creation of small to medium- 
sized digital collections. 

A course entitled "Introduction to Digital Imaging" was first 
taught at the Queens College Graduate School of Library & 
Information Studies (GSLIS), City University of New York (CUNY), in 
the Fall of 2003. In the Spring of 2005, a year-long METRO-funded 
grant facilitated a co-operative project between the Rosenthal Library 
and the GSLIS to support student digitization of a portion of the 
Queens College Rosenthal Library Archives (e.g. see GSLIS, 2005- 
2009, Digitization projects). The project included a variety of forms 
and formats. The evaluation of this valuable learning experience 
identified a strong need to find a single standard format that was 
information rich and moderate in scope, but which lent itself to more 
uniform metadata standards and digital specifications. The evolving 
project, "Waterways of New York", an online digital collection of 
historical postcards, was created in 2006, and partially supported by 
METRO through continued access to CONTENTdm. It continues to 
be extended by GSLIS students each semester the course is taught. 



Scope and Format 

The most important feedback provided to our team by METRO 
digitization experts regarding our "Rosenthal Library Archives" 
initiative was the value of working with a limited number of 
manageable formats and a relatively focused subject area and time 



102 Waterways of New York Project 



frame. During the implementation of the grant a serious problem was 
the complexity resulting from too many different types of media, the 
overly wide range of subject matter, and the challenges these 
characteristics presented to the creation of consistent metadata. 

One of the GSLIS professors, Thomas Surprenant, has an ever- 
expanding collection of Erie Canal and related New York State 
waterways antique postcards, which addressed many of the problems 
noted in the METRO feedback. In particular, by selecting a single, 
simple, information rich format— postcards published before 1923— 
copyright concerns were eliminated and only a single set of 
digitization specifications needed to be developed. METRO'S 
willingness to host the collection on their CONTENTdm server 
simplified selection of Dublin Core as the metadata standard, and use 
of a subset of the Library of Congress Thesaurus of Graphic Materials 
(TGM) for standardized metadata terminology (Library of Congress, 
2007). This greatly aided our ability to develop a manageable set of 
project-specific guidelines that could be adequately addressed by the 
evolving documentation. 

The choice of postcards as the source medium turned out to be far 
more interesting to the students than was expected. An initial option 
to describe the backs—as well as the front images of cards— was 
enthusiastically embraced by virtually all of the students and became 
the norm for subsequent classes. Hand-written messages, address 
conventions, postmarks, trademarks and other attributes of the cards 
were at times as much or even more rewarding to analyze than the 
front images themselves. Further, student interest in the varied 
aspects of architecture and activities of daily living portrayed in the 
postcards led to an expansion of emphasis far beyond the initial focus 
of the project on locks, canal boats, shipping, waterways and 
transportation. 

Background, Technical issues and 
Implications for Teaching 

Any planning for digitization requires a detailed analysis of one's 
institution, and an assessment of where the proposed project fits into 



Waterways of New York Project 103 



its mission and priorities. Further, consideration of the potential 
audience(s), project goals and objectives, resources and limitations, 
oversight and long-term maintenance are among the many issues to 
be addressed (e.g. see JISC Digital Media, 2008: Project management; 
North Carolina Echo Project, 2007). These considerations inevitably 
will shape the nature of the evolving project. It is important that an 
honest appraisal be conducted, committed to writing, and approved 
by the appropriate governing bodies. However, the nature of digital 
projects ensures that adjustments inevitably will be required over 
time. Changing standards, software and hardware upgrades, technical 
glitches, and shifts in the growth of a project are just a few of the 
issues which must be dealt with, often on very short notice. Planning 
and documentation therefore should be viewed as an iterative process, 
where ongoing evaluation is used to address and correct for changing 
circumstances. 

Creating a list of stakeholders and intimately involving them in 
this planning process is critical to success. In our own case, student 
feedback on procedures and emphasis has been an invaluable aspect 
of the evolving project. Each incoming class section serves as a de 
facto Advisory/ Editorial Board that contributes to the decision- 
making process. These contributions include identification of 
additional TGM terms for our thesaurus, the development of 
standardized Trademark descriptions, fine-tuning of documentation 
and lab handouts, and increasingly higher expectations for the quality 
of the metadata. Within a more traditional library environment, all 
members of the digitization team, as well as users and other staff 
members, undoubtedly will have many valuable insights to contribute. 

Among the key elements shaping the evolution of a project-based 
digitization course at Queens College were the following: 

Institutional characteristics 

• When the initial course was developed it was necessary to have the 
course proposal cleared with the departmental Curriculum 
Committee after consultation with the Chair. This required the 
development of course goals and objectives, specific readings, and 
course assignments and activities. 



104 Waterways of New York Project 



• After three semesters teaching the course it was submitted to the 
GSLIS and College Graduate Curriculum Committees, Faculty 
Senate and, ultimately, the CUNY Board of Trustees for approval 
as a permanent course. 

• An understanding of the possible pitfalls of the process at every 
step was important to ensure that all potential hurdles were 
considered and cleared. 

• Even outside of explicitly academic environments, proper 
attention to obtaining documented approvals and support from 
key stakeholders—at all levels up to the governing board— will 
prove invaluable in avoiding challenges and ensuring continued 
buy-in by the institution and other funding agencies. 

Lab facilities (capabilities and challenges) 

• For our project we were able to use a 16 workstation Mac lab with 
direct connections to the Internet. The lab had been expressly 
designed by GSLIS faculty for digitization-related activities and 
hands-on learning, in close collaboration with the Queens College 
Office of Converging Technologies (OCT) and college architectural 
staff, in conjunction with the development of the course proposal. 
Appropriate institutional commitment to fund, support and 
regularly upgrade such a lab was, and is, essential to the 
continuing success of the project. 

• Specifications included an instructor workstation (in addition to 
student Macs), ceiling mounted projector and wall screen for 
demonstrations, two (eventually three) flatbed scanners, 
SilverFast AI scanning software, Photoshop, and the Microsoft 
Office Suite, particularly Excel. 

• A major continuing challenge concerns computer and software 
upgrades. The OCT staff do not always consider the rhythms of 
the academic year in making changes to the lab, which regularly 
causes problems, even after many years of teaching the course. 
For example, in the Fall 2009 semester alone, new computers 
were installed during the first week of the semester. This resulted 
in equipment and software glitches, and a delay in the availability 



Waterways of New York Project 105 



of the lab, as well as the need to test software functionality and 
then revise/upgrade lab handouts with minimal advance notice. 

• While we were grateful for the new equipment (a regular 
replacement cycle is essential for ongoing functionality), timing 
issues resulted in a rough start to the semester. 

• A major equipment problem for us was solved when Apple 
changed to an Intel CPU. The new Mac computers are now dual 
boot (Apple and PC Operating Systems), meaning that they can 
now run CONTENTdm (CDM) using the Project Client software 
interface. Previously, lab sessions had to be specially scheduled in 
a nearby PC lab (CDM Project Client software is not available for 
the Mac OS). However, dual boot capabilities have created 
additional problems of compatibility, accessibility and ongoing 
troubleshooting. 

Software 

• As noted above, new equipment means software installation and 
the attendant complications. The specialized nature of our lab, and 
lack of teaching assistants, necessitates that course faculty test all 
functionalities and work with OCT staff to address problems. 
Oftentimes this has meant repeated testing and troubleshooting, 
frequently a day or two prior to a scheduled class. Such technical 
malfunctions can wreak havoc on the best-planned teaching 
schedule. 

• While CDM has been sufficient for our needs, and we are 
extremely grateful to METRO for their continuing support, there 
are still some issues that cause concern. The biggest is that 
students cannot directly upload their input into the database due 
to the administrative rights structure. This situation requires 
another level of review by the course instructors serving as 
database administrators/quality control experts, adding 
substantially to time demands near the end of the semester. In 
addition, after students submit their data entries for approval, 
editing ability on their side is extremely constrained, by both time 
and software limitations. 



106 Waterways of New York Project 

• More recent upgrades appear to have adjusted this limitation, 
permitting downloading of materials from the live database for 
additional editing if errors are detected. However, this creates 
additional levels of oversight and complexity, and assumes that 
the instructors will be able to approve the uploads in time for the 
students to review and make changes. This is simply not readily 
accommodated within a 15 week course schedule. 

• Further, although recent versions have been more stable, in the 
past CONTENTdm has crashed frequently, causing much 
frustration on the part of both students and faculty. 

• These points emphasize the steep, and ongoing, learning curve of 
digital project-based courses for faculty, support staff and 
students. 

Support 

• Adequate and timely support for equipment and software is 
essential to any technology-based project. The GSLIS has a 
number of student computer assistants and a campus-wide Help 
Desk, but as noted, the specialized nature of our lab sometimes 
puts it outside of the realm of their expertise. 

• It is good practice to fully document and save ALL help desk 
requests and related support communication. These include 
emails, screen shots and help desk tickets. These records of 
ongoing and recurring problems have proved to be invaluable in 
our efforts to ensure follow-through, and to support our case when 
requests have not been fully resolved to our satisfaction or when 
problems repeat themselves. 

• CONTENTdm Help seems to work best when we go through 
METRO. That means that an additional layer of contact needs to 
be activated anytime is a problem. That said, all relevant staff at 
METRO, over the years of this project, have been incredibly 
knowledgeable, supportive and responsive to our needs. 



Waterways of New York Project 107 



Staffing/oversight 

• Experience suggests that having a subject expert for image content 
is a critical factor. The faculty have, or have developed over time, 
sufficient expertise to assist students in their metadata and 
description activities. 

• Given the need to protect the postcards and the equipment the lab 
has to be under supervision whenever anyone is working. This 
greatly adds to the time burden of both faculty and staff. 

• Postcards are stored in archival quality sleeves and students use 
white gloves when handling the postcards while scanning. 

Class size and student characteristics 

• The class size is dictated by two elements: the number of work 
stations and the work volume. Experience suggests that all 
students need to have access to their own workstations, and two 
workstations are dedicated to scanning use (a third must be 
shared between functions). The initial classes scanned, created 
metadata and submitted for approval six cards (front and back), 
but many of these initial canal cards were fairly simple rural 
scenes. Given the amount of detail that has emerged in postcards 
in later semesters, we have gradually reduced input to three cards 
(front and back), because the quality and quantity of the metadata 
has increased substantially. The time spent in quality control by 
instructors has increased commeasurably, despite having students 
doing quality control on their partners' work. 

• Those involved in the digitization process are best served with at 
least an intermediate level of computer and software expertise. 
We have constructed our teaching labs in a step-by-step fashion, 
and utilize in-class time extensively. This allows the faculty to 
introduce and demonstrate skills and to detail the various steps of 
the project. 

• In our experience, students who are highly competent in 
computers and/or relevant software or metadata creation have 
been more than eager to assist their classmates. This leads to a 



io8 Waterways of New York Project 



highly supportive class environment in which all learn from one 
another, modeling (one hopes) the ideal workplace environment. 

• However, a fair number of students have no previous familiarity 
with the Mac OS, Photoshop, scanning and related software, which 
complicates the pace at which the class can proceed. The nature 
of our curriculum and scheduling constraints make it difficult to 
require pre-requisites beyond the required core courses. 
Consequently, a teaching assistant to help in quality control, and 
in the provision of additional technical support in lab sessions, 
would be extremely desirable for all. 

Evolving nature of the target collection 

• Initially, the Waterways Post Card Collection consisted principally 
of cards of the Erie Canal, with collections of additional New York 
State canals such as the Oswego, Seneca, and Champlain Canals. 
As demand for the course has remained steady, indeed increased, 
the diminishing availability of canal-related cards posed a 
potential problem. (Most cards are currently obtained on eBay, 
and changing availability in geographic scope is an interesting 
topic for another paper.) On the other hand, many of those canal- 
related cards that have become available are increasingly 
distinctive. 

• With the Quadricentennial Celebration of the discovery of the 
Hudson River approaching in Fall 2009, it appeared to be a logical 
extension of our collecting scope to extend to another key New 
York State waterway, the Hudson River. We included in our 
selection criteria cards depicting New York Harbor and the East 
River. The first such Hudson River cards were digitized in Fall 
2008. 

• The expansion in scope of the cards created fascinating but 
unanticipated challenges. Publishers, trademarks, and the 
increasing complexity of the images depicted required a 
substantial expansion of the TGM thesaurus, as well as the 
development of descriptions of an increasingly diverse set of 
trademarks, logos, stamp boxes and postmarks. 



Waterways of New York Proj ect 109 

Consistency and accuracy 

• Digitizing any collection over a period of time by a changing group 
of participants creates somewhat greater consistency and accuracy 
problems than might apply in a short-term. In spite of the 
iterative editing of documentation, inconsistencies and errors are 
regularly emerging in our project. To a large degree, this is due to 
the pressures of a fast-paced curriculum, a constantly changing 
panoply of operating systems, software versions, additions to our 
thesaurus and the ongoing, changing nature of the cards within 
our purview.. 

• Our experience has shown us that the students themselves are the 
best editors in catching errors and inconsistencies. It is obvious 
that the road to "perfect" metadata, documentation, labs and 
handouts is continuous, difficult, and perhaps, ultimately elusive. 
Such is the nature of a work in progress. 

• It helps to have students who have a keen eye for detail as well. 
The project was significantly enhanced when then-student Susan 
Savage completed an Independent Study project in Spring 2007, 
that corrected many of our past mistakes, and developed the 
scaffolding of our current metadata documentation. It is now 
obvious that outside help in editing is an important part of the 
process (although not easily achieved). 

Key readings and course activities and relevance for the project 

• Clearly, carefully selected readings are critical to the success of a 
digitization project. We provide access to a range of resources in 
an effort to meet the needs of those at varying levels of familiarity 
with digitization and related issues. Alumni feedback has 
suggested the importance to many of providing continued access 
to the resources once our graduates are working in the field. Once 
involved in a real world project, many become even more aware of 
the importance of items that may not have seemed salient at the 
time of the course. 

• Students complete a "Tech Review Exercise" in week seven, to 
document their understanding of key technical concepts. This 



no Waterways of New York Project 



reduces the need to cover many basic concepts during lectures, 
and to focus on practical development of skills and discussion. 

• In addition, students prepare a detailed case study of a well- 
documented digital collection, and present an overview of their 
sites to class members. Examining the successes and challenges of 
a substantive collection reinforces the concepts and principles 
learned and applied throughout the course. Students comment 
with great pride on the degree of detail of the metadata they have 
developed for their cards, as compared to some of the case studies 
examined, including many well-funded projects. For many 
respondents to our end-of-semester evaluation surveys, the case 
study is perceived as a high point of the course. 

Project Procedures and Workflow 
Assignment of cards to students 

• Postcards are carefully selected in order to give each student a 
maximum of exposure to bibliographic richness, varying formats 
and levels of difficulty. For instance, at least one card will have 
writing on it for transcription. Cards also are selected on the basis 
of color or black & white, as well as postcard era (e.g. divided back; 
undivided back). In addition, postmarks, publishing information, 
trademarks and stamps are considered in allocating cards. A 
lecture on the history of postcards provides background in 
understanding the evolution of these standards (e.g. see 
Smithsonian Institution, n.d.). 

• Once selected, each card is given an eight alphanumeric accession 
number that identifies the subcollection, card number (a total of 
9999 entries are available) in that series, the type of image (access 
or archival), and whether or not it is a front or back view. Early on 
in the project it was realized that having a related but unique 
identifier for each side was vital to having both sides of a card 
displayed together. At this point the Waterways series identifiers 
are: Champlain (c), Erie (e), Delaware and Hudson (d), Hudson 
River (h), and Seneca (s). Thus, a card in the Hudson River series 



Waterways of New York Project 111 



available on the Web will have two accession numbers - one for 
the front and one for the back: e.g. hoo62aci, hoo62ac2. Slightly 
different filing naming schemes are employed for the master 
archival files (e.g. h0062ar1.tif), so that the nature of the file 
format is evident even without the file extension. 

• We adhere to the ISO 9660 8.3 alphanumeric naming standard to 
ensure that our files will be compatible across platforms (JISC, 
2008, Choosing a file name). 

• Stickers with the student names and the accession numbers are 
printed and attached to the outside of the archival sleeves that 
contain the assigned postcards. The sleeves are put into a binder 
so that students can access them, and photocopies of fronts and 
backs created to ensure a record of assignments. 

Scanning and Creating Derivatives 

• Once the postcards are distributed to students, they engage in 
hands-on instructional labs in scanning. Students use SilverFast 
AI software, in conjunction with the latest version of Photoshop, 
using the currently available flatbed scanners (these are regularly 
updated) to create uncompressed Tagged Image File Format 
(TIFF) scans. The resulting archival files, in the 20Mb range, are 
stored (and backed up) off line and can be examined by students if 
extensive detail is needed. These files are the heart of any 
database because they can be used for non-web purposes 

• Using Photoshop, students then create JPEG derivatives at 
medium compression with a resolution of 150 ppi with 1000 pixels 
on the long dimension (following the CDL Guidelines for Digital 
Images: California Digital Library, 2009). The "ar" on the archival 
files are changed to "ac" on the accession numbers to reflect the 
change in file size and specifications. In this process the archival 
masters are reduced in size to approximately 200 KB, appropriate 
for web viewing, and yet still large enough to zoom for some 
degree of detail. 



112 Waterways of New York Project 



Assigning Metadata 

• The heart of the course is when students assign metadata. A group 
exercise using different sample postcards for each pair of students, 
which is then discussed within the class, provides basic experience 
in assigning subject terms. What is initially thought of as a simple 
process quickly becomes complex when the students are 
confronted with the reality of their individual cards. 

• Each student is given a thesaurus, based primarily on the 
Thesaurus of Graphic Materials (TGM), that contains all of the 
subject terms to be used; an electronic versions also can be used to 
copy/paste the terms. In addition, detailed handouts with 
trademark, stamp data, and metadata procedural guidelines are 
made available in print and digital form. Metadata and JPEG 
images will be inserted into an Excel template, which lists all of 
the metadata requirements according to CDM fields (tailored to 
our project), and the corresponding Dublin Core fields. 

• As needed, additional terms from the TGM are added to our 
project thesaurus to reflect characteristics not previously 
encountered. For example, the move to using cards depicting New 
York Harbor required the addition of such terms as skyscrapers, 
aquariums and ferries. 

• Each week, short lectures introduce the students to canal history, 
lore and terminology. Once the basics are covered students are 
then given time to insert JPEGS of the fronts and backs o f their 
first card (the simplest) into an Excel template. They then proceed 
to select subject terms, develop descriptions, and identify key 
Dublin Core fields (e.g. Title, Creator, etc.). 

• The faculty provides assistance and guidance during this process, 
and students are encouraged to work with a partner. Initially 
there is a great deal of trepidation, but as students gain experience 
and confidence less faculty attention becomes necessary. 



Waterways of New York Project 113 

Quality control teams 

• In order to provide more experience to the students, and to serve 
as a double check before the metadata is passed on to the faculty, 
they are paired up with a quality control partner, and required to 
exchange their Excel metadata files. Each student then reviews 
his/her partner's metadata, starting with the initial card, and 
makes any necessary editorial changes and/or suggestions for 
improvement using a different colored font. The reviewed 
metadata set is then forwarded to the faculty for their comments 
and edits, added in yet a different color font. This provides an 
iterative record of changes: the input of initial author and control 
partner, and final corrections by faculty. After receipt of faculty 
feedback, the teams can then proceed with subsequent cards. 

• This staggered approach is recommended because students tend 
to make the most errors in their first attempts, and to learn from 
the ongoing feedback. Additional benefits to the project include 
the identification of new trademarks in the initial set of cards, 
descriptions of which can then be shared with others on 
subsequent cards. Lastly, repeated mistakes may reveal 
unintended errors or inconsistencies in documentation that can 
then be revised to reduce future errors. 

• While this would seem to be a straightforward process it has 
turned out to be much more challenging than first envisioned. 
Yes, the students get, essentially, experience with six, not three 
cards which is good. But a number of other problems have 
resulted. It is almost always the case that the better students pair 
up and that they turn in metadata that needs little comment. 
However, it still takes an average of fifteen minutes per relatively 
error-free file for faculty to review submissions. Weaker students 
require much more faculty review time either because the editing 
process is sloppy or submissions are late. In some cases faculty 
review and editing can take as much as an hour per Excel file. 
Given time constraints and course deadlines this can get extremely 
stressful to all parties. 



114 Waterways of New York Project 

• That said, it continues to surprise the faculty that: l) the quality of 
most metadata submissions is so impressive, and that 2) there 
remain previously unidentified errors in what seems to be a fairly 
strict process. In this regard the quality control process is working 
as envisioned. 

Below is a transcribed example of the metadata for "1609 • HUDSON-FULTON 
CELEBRATION • 1909 [front caption] (Ifront) [h0189ac1]" after it has been uploaded 
to CDM. 



Title: 1609 HUDSON-FULTON CELEBRATION 1909 [front caption] (Ifront) 
[h0189ad) 

Creator: Copyright 1909 J. Koehler, N.Y. [indicated on front only] 

Subject — Front: Cliffs, Clouds, Flags, Grasses, Portraits, Rocks, Ropes, 
Schedules (Time plans), Ship equipment & rigging, Shrubs, Smoke, 
Smokestacks, Steam engines, Trees, Men, Passengers, People, Color 
postcards, Sailing ships, Side wheelers, Aerial views , Rivers. 

Description-Front: A commemorative postcard celebrating the 300 th anniversary 
of the discovery of the Hudson River, with portraits of Henry Hudson and 
Robert Fulton superimposed over a daytime aerial view of the Hudson River. 
Prominently featured are a sailing ship (circa 1609) and steamship (circa 
1809) [presumably the Claremont] which together serve to commemorate the 
passage of time from discovery to the modern day. Soaring cliffs line the far 
bank and along the near bank; at right, there is a gathering of people (perhaps 
Native Americans). An information box titled 1609 HUDSON-FULTON 
CELEBRATION 1909, lists the following 15 events: Sept. 25 Commencement 
Day N.Y., Sept. 26 Religious Observance Day NY., Sept. 27 Reception Day 
N.Y., Sept. 28 Historical Parade NY., Sept. 29 Commemoration Day NY., 
Sept. 30 Military Parade Day NY., Oct 1 Naval Parade N.Y., Oct 2 Naval 
Carnival Parade N.Y., Oct 3. Religious Day Upper Hudson, Oct. 4 Dutchess 
Co. Day, Oct. 5, Ulster Co. Day, Oct. 6 Green Co. Day, Oct. 7 Columbia Co. 
Day, Oct. 8 Albany Co. Day, Oct. 9 Rensselaer Co. Day. COPYRIGHT 1909 
BY J. KOEHLER, N.Y. [indicated on front only]. 

Coverage - Geographic: Hudson River, New York and New Jersey 

Date Original: 1909? 

Publisher: Graduate School of Library and Information Studies - Queens College 
(CUNY), New York, New York 

Language: eng 

Source Height: 3.5" 

Source Width: 5.5" 

Source: Waterways Post Card Collection of Thomas T. Surprenant: Hudson River 
Type: Text; Image 



Waterways of New York Project 115 



Digitization Specifications: Archive masters were scanned at 600 ppi, 24 bit 
color using a Canon CanoScan 88800F and SilverFast Ai scanning software 
and saved as uncompressed TIFFs using Photoshop CS$ and Mac OC 10.5, 
Derivative access files were saved as medium quality JPEGs, 150 dpi, 1000 
pixels on the long dimension. Thumbnails were generated automatically by 
CONTENT dm. 

Date Digital: 2009-09-30 

Resource Identifier: h0189ac1.jpg 

Format: Image 

Format [Medium]: Postcards 

File Size: 218972 Bytes 

Checksum: 287714794 

Height (of Digital Image): 1000 

Width (of Digital Image): 637 

Color Space: sRGB 

Rights Management: Contact Thomas T. Surprenant, ebeltoms@nyc.rr.com 

Contributing Organization: Graduate School of Library and Information Studies, 
Queens College (CUNY), New York, New York 

Digitization Team: GSLIS 757, Digital Imaging Fall 2009,: Beth Daniel Lindsay, 
Meg Donabedian 

(GSLIS, 2006-2009, Waterways of New York V. 2. 1). 



Uploading to CONTENTdm 

• By mid-semester, the students are introduced to the basics of the 
CDM Project Client software in a hands-on lab. They upload a 
sample image, are guided through the creation of a "Waterways of 
New York" banner to be automatically inserted for future uploads, 
and— following step-by-step directions— use the Template Creator 
function to input recurring fields (e.g. Source, Digitization 
Specifications, Digitization Team), and to enable CDM's capacity 
to automatically capture technical metadata such as Resource 
Identifier, File Size, Format and Checksum fields. 

• They then upload another sample image to demonstrate that these 
functions are working, and to doublecheck (and correct for) for 
possible errors in data entry. 



n6 Waterways of New York Project 



• The following week (assuming appropriate progress on early steps 
in the workflow), they are ready to upload their first postcard 
JPEG images (front and back, entered as individual files). Early on 
in the project we decided against uploading the fronts and backs of 
cards as compound objects, since this would severely limit the 
amount of detailed metadata that could be made available. 

• Students "cut-and-paste" the doubly reviewed metadata from their 
Excel files into the appropriate fields in CDM. This approach was 
especially important when we had limited access to the Project 
Client interface in the adjacent PC lab, and at times of instability 
in CDM at the client end. 

• Following an additional review by faculty for completeness and 
accuracy at the desktop, students then upload the files to be 
approved in the Administrative interface. 

• It provides tremendous positive feedback to all to see the fruits of 
many weeks of work appear live in the evolving database, in real 
time, during this most momentous class session. 

• In subsequent weeks, a portion of each class is devoted to 
additional project work, reflecting whatever particular step each 
student is addressing. 

Timing /scheduling issues 

• Given the pressure of a 15 week semester, and the need to respond 
quickly to student submissions, the entire quality control process 
takes much more time than initially projected. Yet, there seems to 
be no alternative to preventing major errors from creeping into the 
database. And students regularly report, in mid-year and year-end 
evaluations, that they find the hands-on nature of metadata 
creation and feedback to be among the best aspects of the course. 

• The time constraints on classes (once a week for 2 hours 35 
minutes) make it difficult to teach what is needed and to allow 
students the time to practice. This is particularly true for 
commuting students, most of whom hold full-time jobs and may 
come to school only once a week. This is complicated by 



Waterways of New York Project 117 



extremely time-consuming commutes (for students living in 
Brooklyn, three or more hours round trip is common). 

• It is not unusual for a few students to still be attempting to scan 
(or rescan) their postcards during the seventh or eighth week of 
class (or to have to redo their derivative files due to lack of 
attention to digital specifications). This puts the faculty— and 
student partners—into a real bind when it comes their turn to 
evaluate student work. 

How much detail is enough/too much? 

• This question has vexed both the faculty and students since the 
inception of the course. One of the best aspects of the project is 
the information rich content of the postcards, both back and front. 
Students have been particularly diligent in finding, and fighting 
for, new subject terms and descriptive material. This makes the 
course an ever evolving environment with an endless series of new 
things to consider and debate. 

• Especially dedicated students have taken the time to research 
historical events depicted in their cards, leading us to add a 
"References" field in Fall 2008, where they can add more 
background and even Web citations. 

• Student interests, skills, and insights have made this very much a 
"student- driven" course, in the best sense of the phrase. Every 
class has made unique contributions to the quality of the project, 
and all (both faculty and students) learn from one another. The 
collective contributions of the myriad participants in this project 
cannot be understated. 

Supervisory issues 

• As previously noted, the class inevitably encounters difficulties 
with students who either are forging ahead or are lagging behind 
others. In any given class the faculty can expect to see a certain 
level of frustration as students cope with their individual abilities 
and challenges. 



n8 Waterways of New York Project 



• The evolving documentation, hardware and software problems 
sometimes create an illusion that the faculty are not well prepared. 
It doesn't take much to derail what has been planned for any given 
class. This adds to the timing and scheduling issues discussed 
above. On the other hand, as more perceptive students note, this 
is the way projects function in the real world, and provides 
insights they can share in future job interviews of how they dealt 
with unforeseen complications. 

Role of independent study students in quality control review 
and overhaul of documentation 

• To date, there have been only two Independent Study students 
interested in assisting with quality control, one for the METRO- 
funded archival project (Amy Armstrong), and one for the 
Waterways project (Susan Savage). They both made extraordinary 
contributions to the projects in terms of documentation and 
quality control. There is no doubt that, in the near future, more 
students and/ or faculty will have to be brought into the process of 
oversight and review in order for existing problems and errors to 
be addressed, but there are substantial challenges in envisioning 
how this may be possible in the current economic climate. 

• The same difficulties exist with the revision of documentation. 
Incremental changes have resulted in some inconsistencies, and 
outright errors, in the documentation. At this point having better 
documentation has a higher priority than quality control review of 
the existing database. Better documentation will result in better 
quality control. During the coming year the faculty intend to make 
major revisions of the documentation to ensure that it is not only 
up-to-date but more accurate. 

Expansion of the database 

• We anticipate that, sometime in the near future, the database will 
have to expand beyond a personal collection. Two preliminary 
moves in that direction have already been undertaken. The 
Ellenville Public Library and Museum permitted its Delaware and 
Hudson (D&H) postcard collection of 70+ cards to be scanned and 



Waterways of New York Project 119 



inserted into a separate Greenstone database. Similarly, the Erie 
Canal Museum in Syracuse permitted us to create archival Tiffs of 
55 postcards. Students then created derivatives and assigned 
metadata on these cards, in addition to several they scanned 
themselves. 

• From experience it is, indeed, gratifying to see that a spirit of 
participation exists within many of the various libraries, museums 
and historical societies. More needs to be accomplished in this 
arena and there are plans to work with additional organizations in 
the coming year. 

• Once the collection becomes more mature, enlisting other 
postcard collectors to contribute to the database is a strong 
possibility. 

• Any future expansion will have to consider moving beyond 
postcards into other forms and formats. While this is a logical 
extension, especially with photographs, maps and important 
historical documentation, the level of difficulty increases. In order 
to keep students in the process it is likely that the faculty will have 
to do a great deal of traveling to digitize collections that are all 
over the state. 

Successes and Lessons Learned 

The literature of "Best Practices" emphasizes the importance of an 
"Advisory Council", and the involvement of key stakeholders, to 
provide feedback and insight into the successes and challenges of an 
evolving project (e.g. Chapman, 2000). Our students have exceeded 
every expectation in this regard, and reinforce the importance of input 
from a variety of stakeholders in the advisory process. Subject 
headings and description have expanded from an initial focus on 
canal-specific characteristics (e.g. tugboats, locks), to many attributes 
of the environment and daily life in early 20th century New York. 
Detailed rules and guidelines provide standardized descriptions 
(developed by the students) of trademarks, logos and postmarks and 
ensure consistency with AACR2 and other standards. 



120 Waterways of New York Project 

Perhaps most importantly, students have gained an appreciation 
for every aspect of the overall planning, processes and implementation 
pertaining to the creation of at least one type of digital collection 
(images). This practical experience is embedded in an exposure to key 
readings in the literature, class discussions, and a detailed case study 
analysis of an existing collection. In particular, they come to 
understand the enormous demands of metadata creation and quality 
control, and the tradeoffs that may lead digital collection managers to 
minimize this aspect of their projects. The role of students as key 
project participants has been extremely valuable in providing concrete 
evidence of their experience as they search for jobs in this difficult 
economy. It has been equally valuable for potential employers seeking 
qualified job candidates. 

At this writing, the project continues, with two course sections 
scheduled for Spring 2010 and no end in sight. Our students have 
gone on to excellent positions in the field and the feedback from 
intern supervisors is consistently positive. With any luck, our 
collection will exceed 1,300 images by early 2010. 

Challenges and next steps 

In addition to many of the future plans outlined above, we are 
seriously considering migration of the database to an open-source 
platform, such as Omeka (http://omeka.org/), Greenstone 
(http://www.greenstone.org/ ) or other available options. Running 
such a digital library collection platform in parallel with CDM would 
present a valuable learning tool for students, while adding to the 
complexity of the project. 

Possible student interest in working on this initiative as an 
independent study project, coupled with a soon-to-be-completed 
search for a new lecturer line at the GSLIS, with specializations in 
digitization and related technologies, provides some basis for 
optimism as we move into the next decade. And we remain 
committed to this labor-of-love that has inspired and empowered so 
many classes of students. 



Waterways of New York Project 121 



Yet on the challenges side, for faculty it remains a difficult 
balancing act. The nature of a project such as this requires continuous 
updating and editing to ensure consistency and quality in the ongoing 
development of the project. This process places serious stresses on 
the competing obligations of supervising faculty for research, 
publication, service, staying current, and responsibilities for other 
courses. This is particularly true during a major period of transition 
within our department. But, some things are very much worth doing, 
and we hope this example may provide encouragement to others 
starting out, or expanding their existing initiatives. We look forward 
to extending the dialogue! 

References 

California Digital Library. (2009, September). "3.5 Guidelines for 
access and thumbnail image files." CDL Guidelines for Digital 
Images, Version 2.0. Retrieved December 31, 2009, from 
http : / / www.cdlib. org/ inside/ diglib / guidelines/bpgimages / 

Chapman, S. (2000). Section III: Considerations for project 

management. In M.K. Sitts (Ed.) Handbook for digital projects: A 
management tool for preservation and access. Retrieved 
December 30, 2009, from 
http : / / www. nedcc.org/resources/pubs. php 

Graduate School of Library and Information Studies. Queens College, 
CUNY. (2005-2009). Digitization projects. Retrieved January 3, 
2010, from http://qcpages.qc.edu/GSLIS/digitization.html 

Graduate School of Library and Information Studies. Queens College, 
CUNY. (2006-2009). Waterways of New York V. 2.1. Retrieved 
January 3, 2009, from http://cdmi2840i.cdmhost. 
com/cdm4/browse.php?CISOROOT=%2Fqcgslis_fo6 

JISC Digital Media. (2008, November 11). Choosing a file name. 
Retrieved September 2, 2009 from http://www.jiscdigitalmedia. 
ac.uk/ crossmedia/ advice/ choosing-a-file-name/ 

JISC Digital Media (2008, November 14). Project management for a 
digitisation project. Retrieved December 31, 2009, from 



122 Waterways of New York Project 



http://www.jiscdigitalmedia.ac.uk/crossmedia/advice/project- 
management-for-a-digitisation-project/ 

Library of Congress. (2007). Thesaurus for graphic materials I. 
Retrieved December 31, 2009, from 
http: / /www.loc.gov/ rr/ print/ tgmi/ 

North Carolina ECHO Project. (2007). Project planning . 

Digitization guidelines, 2007 revised edition. Retrieved January 2, 
2010, from http://www.ncecho.org/dig/guide_iplanning.shtml 

Smithsonian Institution, (n.d.). Greetings from the Smithsonian: A 
postcard history of the Smithsonion Institution. Retrieved 
December 31, 2009, from http://siarchives.si.edu/history/ 
exhibits/ postcard/ chronology.htm 



Part II - A Diverse Digital 
Landscape: Digital Collections 
in Public Libraries, Museums, 
Cultural Heritage Institutions, 
and Knowledge-Based 
Organizations 



Managing Rights in a Medium Scale Audio 
Digitization Project 



Barbara Taranto and Elizabeth Bradley 
(New York Public Library) 



Abstract 

In 2007 New York Public Library received a small grant to digitize a 
selection of audio recordings of public programs, including lectures, 
interviews, and panel discussions that took place at the Humanities 
and Social Sciences Library between 1983 and 2003. This chapter 
briefly discusses the scope of the project and the institutional reasons 
for the initiative. The chapter then discusses some of the significant 
challenges regarding Intellectual Property Rights, faced by the Library 
and the strategies it employed to deal with the issues. The chapter 
concludes with a discussion of the policies and procedures the Library 
has since put in place to manage audio rights. 

Keywords: Access, Audio, Copyright, Media, NYPL, Privacy, 
Public library, Recording, Taping, Transcription. 



The Genesis of the Project 

New York Public Library has been engaged in digital projects since the 
mid 1990's when it launched its first public website. In 1998 it 
published a compilation of digital "reprints" African American 
Women Writers of the i9thCentury (Webpage Dodson, 1998) of out of 
print, and essentially unavailable 19th century materials from the 
holdings of the Schomburg Center For Research In Black Culture. By 
2000 the Library had established a new unit, the Digital Library 



125 



126 Managing Rights in a Medium Scale Audio Digitization Project 

Program, to manage the creation and publication of digitized content 
from its collections. Within the year an even larger project - the 
digitization of half a million pictorial items - NYPL Digital Gallery 
(see New York Public Library Digital Gallery, n.d.) - was undertaken 
and by 2003 the Library was producing approximately 1250 new 
image files - complete with metadata - per week. 

Public domain items were chosen for the first digitization efforts. 
This allowed the Library to leverage existing bibliographic records for 
metadata creation without additional research into intellectual 
property rights. It also allowed the Library to provide a new and 
exciting online resource for the public. 

As the program matured policies and best practices were 
developed, including how to manage the rare but occasional issue of 
Intellectual Property (IP). The metadata system was modified to 
record Intellectual Property rights and access permissions. Processes 
were established to redact items that were mistakenly digitized and/ or 
published before IP permissions had been received, and new content 
was being created daily. At the same time the Library began 
experimenting with new formats such as geospatial data and media 
files. It also began to look closely at user needs and at new avenues for 
distribution such as mobile and cellular applications. 

As patrons became more experienced web users and their home 
computing technology improved, user expectations for easy and quick 
access to rich content increased substantially. Public domain 
materials were no longer sufficient. Consequently in 2007, the Library 
decided to initiate a pilot project to determine what was involved in 
managing rights-encumbered collections in the digital environment. It 
requested and received a grant from the Metropolitan Library 
Association to digitize recordings of well-known public personalities 
and ultimately make them available to the public via the web under 
the title 24 Hours at NYPL. 

The Scope of the Project 

We wanted to test the waters for audio rights management at an 
institution with a mandate for accessible content delivered at no 



Managing Rights in a Medium Scale Audio Digitization Project 127 

charge. There were three key goals: 1) discover the issues involved in 
rights encumbered audio; 2) determine best practices for resolving 
these issues; and 3) develop a workflow to manage collections with 
similar issues. A small sample from a contemporary, but relatively 
high profile collection would be suitable for this purpose since the 
subjects involved were already rights-sawy performers. It was also 
decided that to the best of its ability, the Library would seek 
permission from these rights holders to make these materials freely 
available to the public without cumbersome access controls. 

The Literature 

There is a plethora of documentation on Digital Rights Management 
(DRM) and DRM products available to leverage and control access to 
digital assets for commercial gain and/or Intellectual Property 
management (e.g., ARALOC, 2010, Stream Media, 2010 and Discretix, 
2010.) Everything from watermarking to anti-cloning technologies is 
available to "lock-down" content. Likewise there are as many Digital 
Library resource pages (e.g., Berkeley Digital Library, 2007) detailing 
these technologies and the consequent policy implications for 
managing copyrighted and licensed media in a library setting. 
Furthermore, there is a very lively scholarly discourse in the 
preservation community around the proper expression (language) of 
digital rights and the sharing of digital rights information (see Premis, 
2010). All of which is exceedingly helpful once the data have been 
obtained. 

There are many practices employed when engaged in digital 
activities such as national standards for reformatting, file naming 
conventions, persistence, etc (e.g., Washington State Library, n.d., and 
North Carolina ECHO, n.d.). However, the process and procedures by 
which one should manage a Digital Rights project in a library - 
especially a public library - setting are not thick on the ground. In fact, 
it is for this reason that New York Public chose to work on these 
collections - viz. this is relatively new ground. 

We chose to digitize a selection of audio recordings of public 
programs, including lectures, interviews, and panel discussions that 



128 Managing Rights in a Medium Scale Audio Digitization Project 

took place at the Humanities and Social Sciences Library between 
1983 and 2003. The recordings, holdings from the Public Education 
Program (PEP) represent some of the best of the thousands of public 
programs offered for a small fee to all comers by the Humanities and 
Social Sciences Library during those decades. Many of these events 
were held in a grandly restored space called the Celeste Bartos Forum, 
a venue that has, as a result, become familiar and beloved to 
generations of New York City culture-seekers. 

The programs chosen for digitization included talks and 
interviews featuring world-renowned writers, thinkers, and public 
intellectuals such as Chinua Achebe, A.S. Byatt, Umberto Eco, 
Christopher Hitchens, Robert Thurman, Oliver Sacks, Esmerelda 
Santiago, Sarah Vowell, and Wendy Wasserstein, among many others. 
Permission to record the programs was sought from each subject prior 
to the event. They were originally recorded live, in analog formats 
including reel-to-reel and cassette tapes. 

When choosing these particular programs, preference was given 
to those subjects whose dynamic performances were the most in 
keeping with the goals of the digitization project, and for which 
permission was anticipated to be negotiated and obtained without too 
much difficulty or searching. For these reasons as well, single subjects 
(interviewees) were preferred 

Rights Issues with Recorded Programs 

For the most part flat art works (as plates or reproductions or original 
art works) are associated with a wide but manageable universe of 
rights holders. Even maps which have more "creators" than most print 
materials - engravers, cartographers, drafters, navigators, colorists, 
etc. - are fixed. The rights pertain to certain aspects of the creation of 
the object but do not extend to the subject of the work such as "the 
State of California" or to all parties that participated in the making of 
the work such as the oiler who serviced the printing press. 

The digitally reformatting of the audiotapes was unexceptional. 
The tapes files were sent to a reputable vendor and quality control was 
performed on the files when they were returned to the Library. 



Managing Rights in a Medium Scale Audio Digitization Project 129 

Technical metadata was recorded and deposited into a database 
designed for keeping preservation technical data. The Library's 
metadata specialists using the in-house utility created descriptive 
metadata. However, obtaining permission from the speakers on these 
recordings proved to be an unexpected and ongoing challenge. 

The first step in rights process is to determine who is the rights 
holder. This involved contacting the subjects and in some cases 
editors, managers and lawyers, to match the audio tapes to their rights 
holders. While the subject was clearly known, the rights 
documentation needed to be created (with the assistance of the NYPL 
General Counsel's Office) and managed by NYPL staff, in addition to 
the management of the media asset. In this case, the Library 
deliberately chose items where the rights were clearly held by a known 
body, preferably one with a cordial, ongoing relationship to or affinity 
for the mission and work of the NYPL. Nonetheless seeking and 
obtaining permission even when the rights holder is undisputed is 
often tedious and frustrating. In fact, a good deal of time was spent 
finding and attempting to contact the subject in question. Once 
contacted, there were many issues to resolve. 

Rights holders were often difficult to reach. Sometimes this was 
due to the age of the information (addresses are out of date, names 
have changed, etc) or the amount of time that had elapsed since the 
creation of the content. Sometimes this was due to privacy needs or 
the sheer number of intermediaries (such as editors, agents, and 
business managers) that had to be included in the negotiation process. 
Sometimes there was no one left to contact. The rights may have 
reverted the estate in the case of an individual's death or passed to 
another individual without a clear record. In these cases it was nearly 
impossible to determine the rights. 

In a few cases permission was obtained when considerable time 
was invested. Often rights holders were unfamiliar with the digital 
environment and needed to be educated by Library staff about the 
nature and purpose of digitization. This was a slow process and took a 
good deal of time and patience and to explain everything in adequate 
detail. It also required a finesse to describe technical processes in a 



130 Managing Rights in a Medium Scale Audio Digitization Project 

way that was meaningful to non-technical individuals. Sometimes this 
was successful, sometimes not. 

Certain more commercially successful individuals, whose fame 
had increased in the ensuing years, retained counsel that aggressively 
challenged the Library's request - sometimes with and sometimes 
without cause (e.g., An individual may not want to relinquish rights to 
an unedited version of a performance to avoid the possibility of 
incurring liability for comments made during the performance.) 
These individuals seemed to be motivated by one of two factors - 
control over the distribution of creative content produced at an earlier 
stage in the performer's career, and/or the possibility of generating 
revenue from the digitized program. 

Certain rights holders were unable to remember the contents of 
the programs they had recorded. Others were uncomfortable with the 
idea that their material would be available for free streaming 
broadcast and download via iTunes University. Despite iTunes' 
seemingly universal brand, the prospect of an NYPL partnership with 
a commercial vendor confused and possibly alienated some of the 
performers (or, as was more often the case, their legal counsel). 

One strategic decision made by Library staff was to commission 
written transcripts of the audio recordings of the events that we 
wished to digitize, so that we might make the scripts of the events 
available to anxious authors, publishers and literary agents, many of 
whom suggested that they could not remember the content of a talk 
delivered so many years before, and were therefore reluctant to give 
their approval in advance of reviewing it. 

The issues presented as follows: 

• Rights holder in dispute 

• Rights holder incommunicato 

• Rights holder is the estate of a deceased person 

• Rights holder unfamiliar with technology 

• Rights holder represented by aggressive counsel 

• Rights holder unable to remember contents of discussion 



Managing Rights in a Medium Scale Audio Digitization Project 131 



• Rights holder uncomfortable with venue 

For the most part subjects who participated in these public 
programs were initially hesitant to grant permission to digitize and 
republish the program on the web, which for many of them 
constituted a new and unusual forum for the distribution of their 
work. Interestingly, younger performers gave permission with 
considerably more alacrity and ease than did older performers (or 
their counsel). As a result of these unexpected challenges, only a 
limited number of the digitized public programs have been made 
available to the public. However, the process of negotiating for rights 
and gathering permissions is still ongoing, and over time many more 
subjects have been persuaded to participate in the project. 

The digitized materials are preserved for posterity and the 
existence of this trove of live audio material featuring celebrated 
authors and thinkers (all of whom have book holdings at the Library, 
and some of whom have deposited their personal papers or 
manuscripts, as well) is a source of great interest and enthusiasm for 
Library staff and the users who have learned about them, thus far. 

Lessons Learned 

Few of the digitized tapes have been made available for public access. 
However, the project met two of the three key goals: the issues 
involved in rights encumbered audio were discovered and a workflow 
for handling similar issues was established. The third goal - 
establishing best practices for obtaining permission - can be derived 
from the case studies in the project. 

Many of the subjects were initially reluctant to grant permission 
for very specific reasons. It leads to the conclusion that addressing 
these specific concerns can lead to more successful negotiations. By 
educating the rights holders (and sometimes the counsel) about the 
new medium of digital audio, many of their material concerns were 
allayed, or eliminated. By providing written transcriptions as memory 
aids and evidence to be inspected by counsel, some subjects were able 
to make a positive, informed decision regarding the Library's request. 
By being sensitive to privacy issues, ambassadorial about digital 



132 Managing Rights in a Medium Scale Audio Digitization Project 

content, and above all else, being politely persistent, the Library has 
been increasingly successful. 

New Practices Born From 24 Hours at NYPL 

The entire process of the project is dependent on documentation. 
Without the proper releases in place these projects are not possible. If 
releases do not exist then they must be obtained and rights must be 
clearly spelled out. Furthermore, all records should be managed in a 
centralized repository so that future uses for these items are possible. 
The library was able to identify individual rights holders by sheer 
tenacity and willingness to put in the hours required. The procedure is 
prolonged and often takes many turns. Clear guidelines and processes 
are needed in order to track progress and to properly manage the 
relations established during the negotiations. Without proper record 
keeping the process breaks down and accuracy suffers. Because of 
excellent record keeping the Library was able to pursue this project 
and obtain as many permissions as it has. 

The challenges encountered in obtaining rights for 24 Hours at 
NYPL highlighted the need for the Library to develop a more robust 
rights digital management system. While releases had been obtained 
for the original recordings these releases were paper documents kept 
in separate files. There was no database to manage these documents 
and or the relationship the documents had to the assets. All this work 
had to be done before the work of seeking permission could begin. 
Furthermore, evaluating the releases to determine the extent of the 
Library's rights was necessary before any approach could be made. 

These preparatory tasks are often considered incidental to digital 
projects since digital rights are often discussed as a separate case or 
special case. However, 24 Hours at NYPL underscored the need to tie 
digital rights management to general collection management. In this 
case, the relationship was drawn between the previously obtained 
releases and the audio assets. In the case of the earlier project Digital 
Gallery the issue was obscured because of the decision to include only 
items that were either in the public domain or obtained for the specific 
purpose of being included in the gallery. Nonetheless, the accurate 



Managing Rights in a Medium Scale Audio Digitization Project 133 



documentation in the bibliographic record was the basis for the 
management of the digital rights. 

This understanding has led the Library to create a robust Digital 
Rights Application that will manage scanned paper documents such as 
Deeds of Gift as well as rights information. The data will be managed 
centrally and tied to both the bibliographic databases and the Fedora 
Digital Repository. Rights data that is stored in a structured format 
can be shared with multiple systems and exported for administrative 
purposes and exposed for OAI harvesting. 

Additionally work is being done to educate staff to consider 
digital rights information as vital as provenance to the acquisition 
process; and to include these data in any supporting documentation. 
New guidelines for these managing theses discussions are currently 
being drafted. The focus is to gather as much relevant information as 
possible at the time collections are acquired. Protocols such as those 
listed below are among the suggested practices: 

• Identifying all known individuals contained in items in the 
collection 

• Identifying all known authors and/or creators 

• Identifying all known third party rights holders 

• Identifying permissions and/or releases from third party rights 
holders 

• Identifying documentation related to rights such as releases, 
contracts, etc. 

To the best of its ability the Library is also aiming to secure rights 
as early as possible in the life of a digital object - or in some cases, the 
life of a physical object since this informs almost all future decisions. 
This is a change in practice that is being implemented slowly but 
successfully in the Research Libraries. 

There are many ripple effects to these new practices. In the future 
the decision to acquire or not acquire a collection may rest of the 
status of digital rights. Furthermore, the collection development of the 
entire Library may shift because of thee factors. Certainly, the role of 
the collection development staff will change. 



134 Managing Rights in a Medium Scale Audio Digitization Project 

Summary 

The Library's role as a content creator, provider and distributor in the 
age of podcasts and webinars, wikis, blogs and Twitter is becoming 
critical. Access to new content areas is challenging especially when 
intellectual property rights are involved. However, reluctance by 
rights owners can be overcome by providing written transcriptions as 
memory aids to help make informed choices. Persistence is key. Often 
a donor's point of view can be swayed by gentle persuasion and 
expressed sensitivity to privacy issues. 

Obtaining permissions and releases is essential for the success. If 
not already in hand, the process to obtain these allowances is slow and 
difficult and can impede the completion of the project. 

The lion's share of the work with digital collections with rights 
that obtain to a party that is not the library is the proper recording and 
management of that rights information. All parties involved with these 
collections, including curators, directors, and counsel must share an 
understanding that rights information is not incidental, but essential 
to properly manage these collections. 

Infrastructure - both technical and organizational - that supports 
the recording and management of detailed rights data should be 
implemented in libraries that wish to embark on similar projects. 
Obtaining these data as early on in the acquisition process as possible 
is optimal since researching these data is labor intensive and often 
precludes successful completion of these projects. Staff at all levels 
should be informed of these best practices and educated in the proper 
procedures for acquiring and documenting rights. 

References 

ARALOC. (2010). Cross platform DRM. Retrived on May 2, 2010 
from http:/ / araloc.com/ 

Berkeley Digital Library. (2007). Copyright, intellectual property 
rights, and licensing issues. Retrived on May 2, 2010 from 
http:/ / sunsite.berkeley.edu/ Copyright/ 



Managing Rights in a Medium Scale Audio Digitization Project 135 



Discretix. (2010). Multi-Scheme DRM client. Retrived on May 2, 2010 
from http://www.discretix.com/DRM/index.html?source= 
adwordsCusGdrm-phrase888manufacturersioo2i5&gclid= 
CKqHsMKX3KACFdlw5QodcyfZBQ 

New York Public Library Digital Gallery, (n.d.). Retrieved on May 2, 
2010 from http://digitalgallery.nypl.org. 

North Carolina ECHO. (n.d.). Digitization guidelines. Retrieved on 
May 2, 2010 from http://www.ncecho.org/dig/ 
guide_iplanning. shtml 

Premis. (2010). Preservation metadata maintenance activity. 
Retreived on May 8, 2010 from 
http : / / www.loc.gov/ standards / premis/ 

Stream Media. (2010). Streaming Media Hosting is your expert for 
DRM - Digital Rights Management solutions. Retrieved on May 2, 
2010 from http://www.streamingmediahosting.com/drm.htm? 
gclid=CMLolpWX3KACFWV75QodEAyXDA 

Washington State Library, (n.d.). Digital best practice. Retrieved on 
May 2, 2010 from http://digitalwa.statelib.wa.gov/newsite/ 
proj ectmgmt / vendors .htm 

Webpage Dodson, Howard African American Women Writers of the 
lgth Century. (1998). Retrieved on May 2, 2010 from 
http://digital.nypl.org/schomburg/writers_aa19/intro.html 



The In-House Digital Laboratory: 
Possibilities and Responsibilities 



Andrea Buchner (Center for Jewish History) 



Abstract 

The Gruss Lipper Digital Laboratory at the Center for Jewish History 
has been building sustainable digital collections since 2005. It has 
worked on more than 100 different projects and digitized a wide 
variety of materials. This chapter highlights projects that include the 
digitization of books, photographs, and archival collections. It covers 
digitization from start to finish, touching upon the selection of 
materials for digitization, digitization practices and workflows, the 
management of digital assets, online delivery and users' reactions to 
digital collections. It concludes with a discussion of the feasibility of 
an in-house digital laboratory for archives and libraries in general. 

Keywords: Digital asset management, Digital imaging, Digital 
laboratory, Digital preservation, Digitization, Digitization projects 
management, Digitization standards, Metadata, Sustainable digital 
collections. 



Introduction 

The Gruss Lipper Digital Laboratory at the Center for Jewish History 
was established in 2005 with the generous support of the Gruss 
Lipper Family Foundation. Today, the lab is at the center of a digital 
program whose mission is to preserve the digital assets of the Center 
and its five partner organizations and to ensure long-term access to 
them. Since 2005, the digital lab has been involved in more than 100 



136 



The In-House Digital Laboratory 137 



digitization projects and has created more than 68,000 high-quality 
digital images. It has also created 300 hours of high-quality digital 
audio files by digitizing oral histories and Sephardic music stored on 
cassette tapes and reel-to-reel tapes. 

The digital lab manages all digital assets via ExLibris' DigiTool, a 
digital asset management system for libraries and archives 
(http://www.exlibrisgroup.com). The public interface is CJH Digital 
Collections (digital.cjh.org). The digital collections are an integral part 
of the Center's Online Public Access Catalog (catalog.cjh.org). This 
catalog provides access to the Center's and partners' archival 
collections and library holdings. If a digital reproduction exists in CJH 
Digital Collections, the respective catalog record links to them. As of 
December 2009, 21,000 digital objects were accessible. In 2009 alone, 
174,000 users requested digital objects. All materials are accessible to 
the public free of charge and the collections are constantly growing. 

This chapter will examine a variety of projects. They illustrate 
approaches to the digitization of different types of materials: books, 
photographs, and archival collections. This chapter will also discuss 
digital asset management, online delivery, users' reactions to digital 
collections, and the feasibility of an in-house digital laboratory. 

The Grass Lipper Digital Laboratory 

The Gruss Lipper Digital Laboratory (digital lab) is a department 
within the Center for Jewish History. 

The Center for Jewish History is the home of five prominent 
Jewish institutions dedicated to history, culture, and art. They are the 
American Jewish Historical Society (AJHS), the American Sephardi 
Federation (ASF), the Leo Baeck Institute (LBI), the Yeshiva 
University Museum (YUM), and the YIVO Institute for Jewish 
Research (YIVO). The combined holdings of the partner organizations 
bring together the rich variety of Jewish historical experience, ranging 
from Eastern European Jewry, Sephardic Jewry, German-speaking 
Jewry, to the American Jewish community. 

The collections include more than 500,000 volumes, 25,500 
linear feet of archival documents, and thousands of museum objects. 



138 The In-House Digital Laboratory 



Since its opening in October 2000, over 40,000 researchers have 
visited the Center and roughly 210,000 patrons have attended various 
public exhibitions, conferences, and other events that the Center and 
partners held. 

The Gruss Lipper Digital Laboratory, a department of the Center, 
collaborates with all five partner organizations in numerous ways by 
digitizing their collections, managing the resulting digital assets, and 
providing digital consultancy services. During the first two years of the 
lab's operation it was fully funded through the grant from the Gruss 
Lipper Family Foundation. When this period ended in 2007, the 
Center made the digital lab an integral part of its functions and 
services and continued to offer the lab's digital collections building 
services to the partner organizations. In order to defray some of the 
costs, however, the digital lab charges for its services whenever 
external funding becomes available to the partners. When this is not 
the case, the lab provides pro bono digitization services. 

A typical digitization project starts with an idea put forth either by 
Center or partner staff. This is followed by discussions, examinations 
of materials, and following established procedures (including the 
completion of forms) to ensure that materials arrive in the lab that 
have been thoroughly vetted in regard to their value, physical 
condition, rights, and available metadata. Once the selected materials 
have been prepared and arrive in the lab, the process of digitization 
begins. Projects may involve one or multiple partner organizations. 

Projects 

When it comes to determining the value of materials, the lab works 
closely with partner staff. Value lies in the exceptional coverage of a 
topic, the uniqueness (scarcity) of materials, how well they represent 
the unique strengths of a partner's collection, and their appeal to the 
partner's and the Center's audience. Materials must also be in the 
public domain or be orphan works (i.e., copyrighted works whose 
owner cannot be identified or located. See Orphan Works, 2009), or 
the partner must have permission from copyright holders to make 



The In-House Digital Laboratory 139 



digital reproductions of them accessible online. Preservation needs 
also play a role in selecting materials. 

Among the materials digitized in the lab are 137 Yiddish and 
Hebrew children's books from the Yeshiva University Museum (YUM) 
and the YIVO Institute for Research (YIVO). The books document the 
development of Jewish children's literature from the turn of the 
century until the onset of the Holocaust in Eastern Europe. The 
majority of the books are rare and only available at the Center. They 
are in the public domain or orphan works and could therefore be 
made accessible via CJH Digital Collections (digital.cjh.org). The 
project began as a pilot project funded by the Metropolitan New York 
Library Council (METRO) in 2007 and grew into a larger project when 
a private family foundation decided to fund the digitization of 
additional books. The Center also contributed the digital images to the 
International Children's Digital Library Foundation which has made 
them accessible through its website. 

In addition to children's books, the digital lab has digitized rare 
books from the Leo Baeck Institute (LBI). The digital lab has been 
digitizing rare books since its inception. Funding has come from a 
variety of sources; originally from the Gruss Lipper Family 
Foundation, then from METRO and now through a private donor. 
Most recently, the lab digitized manuscripts from the i5-i6 th century 
pertaining to a famous Renaissance controversy between the Christian 
Hebraist Johannes Reuchlin and the anti-Jewish agitator Johannes 
Pfefferkorn, who advocated for the destruction of all Jewish books. 
The digital lab is ideally suited for the digitization of rare books. It is 
an in-house lab which reduces the risk of loss and damage that can 
occur during the shipping of books to external vendors. Moreover, the 
digitization staff is given the time to treat each book carefully and 
according to its often fragile or tightly-bound condition. This ensures 
that they are not damaged during the digitization process. 

As for photographs, the digital lab has worked on dozens of 
projects during its first two years of operation, having digitized 
thousands of photographs (prints and negatives) as well as slides. 



140 The In-House Digital Laboratory 



Among the highlights are 3,200 photographs of Moroccan and 
Turkish synagogues from the American Sephardi Federation (ASF). 

The digital lab has digitized both portions of archival collections 
as well as complete collections. From the LBI, it has digitized rare 
oversize materials. They were selected due to their rarity and fragility. 
Among the highlights are letters of protection for Jewish communities 
in Germany from the 18 th century. Letters of protection were issued to 
Jews by local authorities in Europe. These letters gave Jews the right 
to reside in a town in exchange for a special tax. 

The digital lab has digitized a few archival collections in their 
entirety. The largest collection was from YUM. It consisted of 30 
linear feet of materials, mostly fashion drawings by the New York City 
fashion designer Abe Grubere. However, this collection is an exception 
and archival collections digitized in the lab are usually small, such as 
the Raphael Lemkin Collection from the American Jewish Historical 
Society (AJHS) that consists of 7.5 linear feet. Raphael Lemkin was a 
scholar who coined the term "Genocide." He was instrumental in the 
United Nations' adoption of the Convention on the Prevention and 
Punishment of the Crime of Genocide in 1948. The collection 
documents Lemkin's lifelong effort to prevent genocide and lobby the 
United Nations to adopt an anti-genocide convention. In total, the lab 
digitized the content of 76 folders which translated into 4033 images. 
The materials were digitized as part of an initiative at the Center that 
culminated in a conference and an exhibition on genocide in 2009. 

Copyright and Fair Use 

When materials are selected for digitization the copyright status must 
be clear. As for the YUM and YIVO children's books and LBI rare 
books, the books were either in the public domain or orphan works. 
They could be made accessible via CJH Digital Collections without 
restrictions. 

With regard to the photographs of Turkish and Moroccan 
synagogues, however, the situation was more complicated. There was 
a subset of 500 photographs of the 3,000 total, taken in 1989, where 
the ASF and the photographer jointly held the copyright. The 



The In-House Digital Laboratory 141 



photographer was then consulted for his permission to digitize the 
photographs and to make the digital reproductions accessible online 
via CJH Digital Collections. His permission was obtained by means of 
a license agreement. 

As for the Raphael Lemkin Collection, the AJHS had permission 
from Lemkin's heir to make digital reproductions of the collection 
accessible. Among the selected materials was much correspondence 
from third parties. They were letters from various activists to people, 
committees or organizations that had the power to influence the 
United Nations or various governments to adopt and/or ratify the 
convention. The letters were protected by copyright. AJHS and the 
Center started to contact copyright holders and permission was 
granted whenever possible. We also reviewed the four factors of "fair 
use" and concluded that the factual nature of the correspondence and 
our intended use allow for making the letters accessible to the public 
for "fair use" (see Fair Use, 2009). Each digital folder containing 
letters was associated with a "click through" copyright notice, 
outlining the allowed terms of "fair use" such as that no use other than 
research, teaching, and private study is allowed without prior 
permission from the copyright owner. Copyright owners not properly 
identified, or any user with information on the copyright status of a 
particular work, are asked to contact the Center and AJHS so that 
appropriate information can be provided in the future. Every user 
desiring access to the digital materials must agree to these terms by 
clicking on a button before gaining access to them. The fair use 
information can also be found in the descriptive metadata record that 
is attached to each letter and at the bottom of the CJH Digital 
Collections website. 

Image Capture 

The Gruss Lipper Digital Laboratory has the following equipment: A 
BetterLight Super 8K-HS digital scanning camera back and a Mamiya 
645ZD 22 megapixel (5,336 x 4,008 pixels) medium format single- 
shot digital camera. The scanning workstation consists of a motorized 
AIAXact 3040 DV copystand, four Buhl HID 150W 4k softcube lights 



142 The In-House Digital Laboratory 



as well as two book cradles. The lab also has an Epson Expression 
10000 XL Photo flatbed scanner with transparency adapter and a 
Minolta DiMage Scan Multi PRO AF-5000 multi-format film scanner. 
The digital lab also has a TEAC 860-R cassette deck, an Otari Mx- 
5050bII reel-to-reel tape machine, and a Behringer UB502 mixer. 
This equipment allows for the digitization of the different types of 
materials found in archives: Bound books, unbound materials, 
oversize materials, photographs, film, slides, transparencies, cassette 
tapes and reel-to-reel tapes. 




Figure GLDL-1: Gloria Machnowski digitizes a decree from the Leo Baeck Institute 
with the BetterLight camera. © 2009. Gruss Upper Digital Laboratory. Used with 
permission. 



When the digital lab receives a collection to be digitized, it tracks 
the collection, item, and images eventually to be created from the item 
in a Microsoft Access "tracking" database. For the actual digitization, 
the digital lab follows standards established by the National Archives 



The In-House Digital Laboratory 143 



(Technical Guidelines for Digitizing Archival Materials for Electronic 
Access: Creation of Production Master Files - Raster Images, 2004) 
which include guidelines for the creation of high-quality archival 
master images, quality assurance and accompanying metadata. As 
these guidelines recommend, each image contains a color scale and 
gray scale as color and tone references and a ruler as a reference for 
the size of the original document. Depending on the original, the 
digital lab produces 400 ppi 8-bit grayscale, 400 ppi 24-bit color, or 
600 ppi l-bit bitonal TIFF images with LZW or ITU-T compression. In 
terms of resolution, photographic prints and film pose an exception. It 
is determined by the size of the original and a minimum number of 
pixels needed to produce a high-quality image. 8-bit grayscale images 
are tagged with a Gray Gamma 2.2 profile, 24-bit color images with an 
Adobe RGB (1998) ICC color profile. 

Recommended technical and administrative metadata that is not 
embedded automatically during the image capture process in the TIFF 
image is added to fields in the TIFF header via Adobe's Photoshop 
software, such as the name and title of digitization staff, scanning 
facility (Gruss Lipper Digital Laboratory), organization (Center for 
Jewish History), and contact information. 

The TIFF images are not post-processed. The goal is to create a 
digital archival master file that reproduces the original as closely as 
possible. During the quality assurance process, the TIFF images are 
inspected for orientation, sharpness, color, contrast, noise, and dust. 
Images failing this inspection (images that are skewed, out of focus, 
over- or underexposed, show digital noise, dirt, or dust) have to be 
rescanned. Access copies for web delivery are created either in the 
JPEG2000 or PDF format. JPEG2000 is a format that has an efficient 
compression rate allowing for the online delivery of high-resolution 
images into which users can zoom to appreciate fine details. This 
format works well with illustrated children's books, rare books and 
oversize materials as well as photographs. At the maximum zoom 
level, the full resolution and detail of the JPEG2000 image can be 
appreciated. 



144 The In-House Digital Laboratory 



For collections consisting of mostly type-written materials, such 
as the Raphael Lemkin Collection, the delivery format is PDF. The 
files are small and load quickly and the full text of each type-written 
document is searchable thereby allowing users to quickly search the 
content of a digitized folder by keyword. 

Metadata 

The Gruss Lipper Digital Laboratory records descriptive, 
administrative, and structural metadata via a variety of community 
standards. It has adopted MARC 21 (MARC Standard, 2010) as the 
standard defining the structure of a catalog record and AACR2 
(Anglo-American Cataloging Rules, 2005) as a content standard. The 
standards are justified since MARC records already exist in the 
Center's library catalog for many of the items digitized in the lab. They 
can easily be imported into the Center's digital asset management 
system using the Z39.50 protocol. The staff of the Center and Center's 
partners consists of trained librarians who update records when 
needed. 

If records are created from scratch and do not yet exist in the 
Center's OPAC, the digital lab recommends DACS as a content 
standard. DACS (see SAA: Describing Archives: A Content Standard, 
n.d.) was not conceived with the space constraints of a library catalog 
card in mind and therefore eliminates many of the abbreviations 
present in AACR2, making it a more user-friendly standard in an 
online environment. 

The lab has used MODS (Metadata Object Description Schema. 
2010) for selected projects, but has not yet officially adopted that 
standard. The lab has developed cataloging guidelines for a MARC 21 
digital collections minimum record and a MARC 21 digital collections 
core record. 

The MARC digital collections minimum and core record are based 
on the idea of levels of description as put forth in DACS, with the 
minimum and core record roughly corresponding to DACS' single- 
level minimum and single-level optimum/added value descriptions. 
Particularly important is the use of controlled vocabulary. Controlled 



The In-House Digital Laboratory 145 



vocabulary allows for the building of browsing categories in the 
Center's digital collections portal, CJH Digital Collections 
(digital.cjh.org). These virtual collections allow users to search for 
relevant materials across collections and partners. Controlled 
vocabulary also ensures that a user retrieves all relevant materials 
when searching by keyword. Subject headings are taken from the 
Library of Congress Subject Headings (LCSH), Thesauri used are the 
Art and Architecture Thesaurus (AAT), and the Thesaurus of Graphic 
Materials (TGM). 

The virtual collections in CJH Digital Collections are based on 
type (the lab uses genre terms to build the "collections by type" 
browsing hierarchy, such as "children's books," "drawings," 
"photographs," etc.), repository (the Center and its five partners), and 
provenance (title of a collection). Most recent research recommends 
organization by subject (Schaffner, 2009). The lab has begun this type 
of organization for a selected set of photographs from YIVO (subjects 
are Holocaust, political life, Yiddish theatre and others). 

A "collection by type" to be highlighted is "archival finding aids." 
CJH Digital Collections hosts 8,200 EAD finding aids, many of them 
converted from MARC as part of a recently completed project. Our 
goal is that eventually all finding aids will link to digital objects, if the 
archival collection has been digitized. These finding aids constitute an 
additional access point to the digital objects. 

Finally, the digital lab also uses OAI-PMH, the Open Archives 
Initiative Protocol for Metadata Harvesting. The use of this protocol 
allows for the harvesting of all MARC records in CJH Digital 
Collections by OAIster, a union catalog of more than 23 million 
records from more than 1,100 contributors. OAIster records are freely 
available through WorldCat.org. 

Creation of MARC Records 

When MARC records already exist in the Center's OPAC, as it is the 
case for all rare books from the LBI, they are imported into the lab's 
digital asset management system and adjusted in terms of genre terms 



146 The In-House Digital Laboratory 

and provenance to ensure they appear within the appropriate 
browsing category in CJH Digital Collections. 

For some of the children's books from YUM, MARC records did 
not exist in the Center's library catalog. YUM, as the only museum 
among the Center's partners, has its own database that is separate 
from the Center's library catalog since museum objects follow 
different rules for description. For these books, a children's book 
MARC record template was created in DigiTool which was used by the 
lab's metadata librarian who cataloged the children's books based on 
the information available in the museum database. For the ASF 
photographs that were digitized by the digital lab, just as with the 
YUM children's books, no MARC records existed in the Center's 
OPAC. In general, not much information was available for them, 
which is not a surprise since photographs that are part of archival 
collections are rarely cataloged on the item level. The little 
information partners usually have about photographs at the point of 
digitization, call number, dimensions, rights and repository that owns 
them, is generally transferred to the lab's tracking database. Once 
digitization is completed, the data is exported from the database and 
converted into MARCXML and together with the images, ingested into 
the Center's digital asset management system. These MARC records 
with minimal description are designed to be enhanced, based on the 
lab's cataloging guidelines by partner catalogers once they have 
obtained the missing information, such as title, date, and other added 
entries. 

In addition to descriptive metadata, the digital lab also creates 
structural metadata. This type of data is necessary for complex objects 
such as books, archival folders, and archival documents that consist of 
more than one page. Structural metadata preserves the physical 
and/or logical structure of such complex/multi-page objects in a 
digital environment. It allows users to page through a faithful 
reproduction of the original online. The digital lab uses METS (see 
METS Metadata encoding and transmission standard, 2010) as the 
standard for structural metadata. 



The In-House Digital Laboratory 147 



The structure or "structural map" as it is known in METS is 
determined by the physical and/or logical structure of a physical item. 
Re-creating structures for books is less complicated and time- 
consuming since pages are usually numbered and chapters exist. 
Much more time-consuming is the recreation of the structure of an 
archival folder. Description usually only exists on the folder level, but 
a structural map for a folder reaches down to the item level. How 
detailed the structural map will be depends on how much description 
is necessary to make the content intelligible to users. 

University archivists reported during the More Product, Less 
Pixels session at the SAA Annual Meeting in 2008 that students were 
frustrated when working with digital archival collections with no item 
level description. In comparison to a physical folder which can be 
opened up quickly and paged through, students felt that it was more 
difficult to determine if the content of a folder was useful to them or 
not. They complained about an "online microfilm experience." (see 
Session 49: More Product, Less Pixels: Alternate Approaches to 
Digitization and Metadata, 2008. ) 

Due to the relatively small size of the Raphael Lemkin Collection, 
the digital lab decided to facilitate users' experience and to describe 
the content of each folder not only on the folder level but also the item 
level. Folder information came from the finding aid but the item level 
information was supplied by the digital lab's metadata librarian. The 
results are extremely detailed structural maps. For example, the 
structural map of a digital folder containing correspondence consists 
of labels for each letter containing the name of the correspondent and 
date of creation, and all letters are organized by date. 

Moreover, additional information that the lab's metadata 
librarian gleaned from studying the letters and particularly Lemkin's 
index cards and research notes was also incorporated. This was done 
by adding subject headings and added entries to the MARC record 
that is attached to each digital folder. Finally, AJHS personnel 
updated the finding aid with links to the digital folders. 



148 The In-House Digital Laboratory 



File Management and Online Display 

The digital lab manages all digital assets via the digital asset 
management system DigiTool, which has been designed for digital 
asset management in libraries and archives. It supports a variety of 
community standards and file formats including audio and video, 
supports complex objects, assigns unique and persistent identifiers to 
each digital image and metadata, supports basic preservation tasks 
such as tracking of changes to objects and metadata, saves this data as 
well as all other metadata in xml, and allows for digital asset 
management en masse, all of which reduces the risk of loss or digital 
obsolescence. The physical infrastructure of the digital repository is 
provided by a StorageTek FlexLine 210R storage unit. This storage 
unit currently contains 10 TB of RAID-5 storage, and a tape library 
system for daily backups to tape. Backup tapes are stored off-site. 
When it comes to storage and backup, the digital lab works very 
closely with the Center's IT department. 

DigiTool, while primarily a digital asset management system, also 
powers the Center's digital collections portal, CJH Digital Collections. 
A variety of viewers allow for the display of all the diverse digital 
assets described in this chapter, such as simple objects like 
photographs as well as complex objects like books and archival 
folders. However, while the viewers are relatively sophisticated 
(offering zooming and rotation features, browsing by a thumbnail 
gallery or a table of contents created through METS), the interface 
cannot be customized beyond color schemes and fonts. External 
applications can be embedded, such as an Adobe Flash player, and 
whenever this is done, the Center's web designer is of great assistance. 
However, one has to keep in mind that DigiTool was primarily 
designed to manage digital assets and not to create exciting web 
exhibitions. 

User Experience 



The digital lab has gained much experience in the creation and 
management of digital assets. Over the last four years, it has built 
sustainable digital collections based on its mission to preserve them 



The In-House Digital Laboratory 149 



and to ensure their accessibility for the long term. However, creation 
and management of sustainable digital assets is only one side of 
digitization, the other is users' reactions to digital collections. Since 
early 2009, DigiTool generates reliable usage statistics which the lab 
started to analyze. At the same time, the lab also joined the photo- 
sharing website Flickr (http://www.flickr.com/photos/center_for_ 
jewish_history) and made user surveys available in CJH Digital 
Collections to obtain feedback. 

Usage statistics for CJH Digital Collections reveal that on average 
in 2009, the site was visited by 14,500 users a month (ca. 470 people a 
day). Of course, certain collections are requested more often than 
others. Among the materials discussed in this chapter, the children's 
book collection ranks the highest in terms of requests, having been 
requested by users between 30-70 times between March - December 
2009. The Raphael Lemkin Collection and LBI rare books rank 
second, with 20-30 requests per folder or book during the same time 
period, with one rare book, "Juedisches Ceremoniel" standing out, 
having been requested more than 78 times. Photographs of Turkish 
and Moroccan Synagogues are among digital objects least requested 
(1-10 times during the same period). The discrepancies in requests can 
be explained by the newness of a project and their availability by 
several means of access. While the children's books and synagogue 
photographs are projects that started three years ago, the children's 
books project was completed only in 2009 when it was incorporated 
into a bigger children's books digitization initiative. The children's 
books are currently featured in a virtual collection dedicated to 
recently completed projects in CJH Digital Collections. They are also 
accessible through the website of the International Children's Digital 
Library Foundation and 14 of them are available on Flickr, as is the 
"Juedisches Ceremoniel." 

The Raphael Lemkin materials are accessible online in their 
entirety only since September 2009. However, the number of requests 
is high due to the publicity they received in conjunction with the 
November 2009 Genocide and Human Experience Conference at the 
Center. In the course of the advertising campaign for this conference, 



150 The In-House Digital Laboratory 

selected materials were featured on the conference's home page and in 
a dedicated web exhibition. 

While the original purpose of joining Flickr was to gain feedback 
from users (DigiTool does not allow for users to leave comments) it 
became clear quickly that Flickr is also a great tool to promote the 
Center and partners' digital collections. The 189 images on Flickr (as 
of 11 December 2009) were viewed more than 12,300 times over the 
course of eight months. This means that on average, an image on 
Flickr has been requested 65 times. This is a much higher average 
number of requests for images in comparison to CJH Digital 
Collections. The lab was also contacted and asked if some images 
could be featured on a personal blog. Moreover, people responded to 
AJHS' requests to identify unknown people in photographs. Among 
the people who offered their help was the great-granddaughter of one 
of the people featured in a photograph. Users tagged images with the 
occasional "that's a wonderful image" remark. There were no negative 
remarks. Considering the feedback and number of requests per image, 
the Center and participating partners deemed the project a success 
and the uploading of images to Flickr and the administration of the 
Center's Flickr site has become part of the digital lab's services. 

The surveys created to obtain users' reactions to digital 
collections in CJH Digital Collections indicate that users are generally 
impressed with the high quality of the images and their delivery. They 
show that users sometimes get frustrated, citing an inability to find 
what they are interested in. Users also find it difficult to get a sense for 
the amount of material online, their content, and how they relate to 
one another. 

From analyzing usage statistics, surveys, and reaching out to 
users through Flickr we have learned that it will become increasingly 
important to better promote digital collections, explain more clearly to 
users what our digital collections are about and how they should be 
used, and connect with users in new ways. 



The In-House Digital Laboratory 151 



Conclusion 

The Gruss Lipper Digital Laboratory represents an impressive 
investment on the part of the Center for Jewish History. During the 
first two years of its operation, a grant paid for the complete 
refurbishment of a room making it suitable for a digital lab, the 
purchase of first-rate digitization equipment, and a staff of five people 
(a director, metadata librarian, photographer, and two part-time 
digital asset production associates). At this point, the Center supports 
a staff consisting of two full-time and two half-time positions: one 
director, one metadata librarian, and two photographers/digital asset 
production associates. There are additional costs for necessary and 
continuous upgrades and investment into software, hardware, and 
data storage as digital collections continue to grow. The cost of 
continued education for staff as digital technologies evolve also has to 
be factored in. Finally, technical support that is lent to a digital lab by 
an institution's IT department must be part of any lab's budget 
considerations. 

Because of these high costs, a state-of-the-art digital lab requires 
sustained support from stakeholders. These stakeholders have to have 
an understanding of the challenges that maintaining a digital program 
brings and an understanding of the basic concepts behind digital 
preservation. With this support and a professional and standards- 
based program for digital image capture and digital management as 
represented by an in-house digital laboratory, the groundwork is laid 
for the building of a trusted digital repository (RLG/OCLC Working 
Group on Digital Archive Attributes, 2002), the long-term goal of any 
digital program. 

For the investment it represents, an in-house digital lab like the 
Gruss Lipper Digital Laboratory brings tremendous benefits. When 
developing new digitization projects an institution can draw on 
experienced staff. The digital assets remain under an institution's 
control, they can be accessed at any time, and there are no access 
restrictions because rights to the digital images have not been given to 
a vendor. The high-quality digital images can be used for many 
different purposes, ranging from web exhibitions to printing and 



152 The In-House Digital Laboratory 

publishing. The professional administration of a digital asset 
management system ensures that digital assets are managed safely 
and securely. No file is lost, even years after a project's completion. 
Strict adherence to community standards ensures that files can be 
migrated to different formats if the need arises and that data can be 
exchanged with other systems. The result is digital collections whose 
hallmarks are usability, portability, and longevity. 

There is much more to digitization than just image capture. It 
includes metadata, standards, digital asset management, online 
delivery and users' experience, all of which require expert knowledge 
for creating useful and sustainable digital collections. If long-term 
preservation of digital assets is part of an institution's mission and if 
that institution remains abreast of new digitization developments and 
technologies, an in-house lab is well worth the investment. 

References 

Anglo-American Cataloging Rules, Second Edition, 2002 Revision, 
2005 Update. (2005) Chicago, IL: American Library Association. 

Fair use (2009). Retrieved December 11, 2009, from U.S. Copyright 
Office website, http://www.copyright.gov/fls/fl102.html 

MARC standards. (2010). Retrieved December 11, 2009, from the 
Library of Congress - Network Development and MARC Standards 
Office website http://www.loc.gov/marc/ 

METS. Metadata encoding and transmission standard (2010). 
Retrieved December 11, 2009, from the official web site 
http : / / www.loc.gov/ standards / mets / 

MODS. Metadata object description schema. (2010). Retrieved 
December 11, 2009, from the official web site http://www.loc.gov/ 
standards/mods/ 

Orphan works: Statement of best practices (June 2009). Retrieved 
December 11, 2009, from Society of American Archivists' website 
http : / / www. archivists, org/ standards / 

RLG/OCLC Working Group on Digital Archive Attributes (2002). 
Trusted digital repositories: Attributes and responsibilities. RLG, 



The In-House Digital Laboratory 153 



May 2002. Retrieved December 11, 2009 http://www.rlg.org/ 
en/page.php?Page_ID=583 

SAA: Describing archives: A content standard (DACS) (n.d.). 
Retrieved on May 1, 2010 from http://www.archivists.org/ 
governance/ standards/dacs.asp 

Schaffher, Jennifer. (2009). Metadata is the interface. Better 

description for better discovery of archives and special collections, 
synthesized from user studies. Retrieved December 11, 2009, from 
the OCLC website http://www.oclc.org/research/publications/ 

Session 49: More product, less pixels: Alternate approaches to 
digitization and metadata. (2008). SAA Annual Meeting, August 
26-30, 2008. 

Technical guidelines for digitizing archival materials for electronic 
access: Creation of production master files - raster images. 
(2004). Retrieved December 11, 2009, from the National Archives 
website http://www.archives.gov/preservation/technical/ 
guidelines.html 



Building a Virtual Library - A Case Study 
at The Library of The Jewish Theological 
Seminary 



Naomi M. Steinberger 
(The Library of The Jewish Theological Seminary) 



Abstract 

The goal of this chapter is to show how a medium-size research library 
with rich resources in special collections has succeeded in the past 
decade, in building a digital library. The chapter focuses on building a 
strategic plan for a digital library and assessing The Library's 
collections for digital readiness. Planning the digital library includes 
making many decisions ranging from software, prioritizing collections 
to digitizing, metadata schemas, and more. Funding for digitization 
can be grants, gifts in kind where the funder provides the institution 
with a digital photographer, to funders who want a specific collection 
digitized. There is a discussion of some of the challenges such as 
hardware capability and staffing. The important lessons learned 
through experience and plans for the future of the digital program are 
discussed. 

Keywords: Assessment, Building digital collections, Digital 
collections, Digital library, Digital project funding, Medium-size 
academic library, Metadata for digital collection, Strategic planning, 
Virtual library. 



154 



A Case Study at The Library of The Jewish Theological Seminary 155 

Background 

The mission of the Jewish Theological Seminary, the intellectual and 
religious center of Conservative Judaism, is twofold: to serve as the 
pre-eminent center for the academic study of Judaism outside of 
Israel, as one of the pre-eminent centers world-wide, and as a training 
center to advance that study; and to educate Jewish professionals and 
lay leaders in the spirit of Conservative Judaism for the total 
community through academic and religious programs, both formal 
and informal. 

In accordance with the Seminary's overall goals, The Library's 
mission is to collect, preserve and make available the literary and 
cultural heritage of the Jewish people. The collection includes 25,000 
rare books, 11,000 manuscripts, 400 archival collections, and many 
other historically significant items. Among the particular strengths of 
its collection are its 35,000 fragments from the Cairo Genizah 
(representing the lives of Jews and others in the eastern 
Mediterranean from the 11th to the 19th centuries), its collection of 
ketubot (Jewish marriage contracts)— the largest in the world— and its 
unparalleled collection of Passover haggadot (the traditional text of 
the holiday, combining narrative and ritual). The Library is also home 
to the world's largest collection of Hebrew incunabula (early books 
printed before 1501). 

The Special Collections are open to students and researchers from 
around the country and the world. Scholars of Jewish history 
specializing in virtually any period or place rely on the 
Library's unique holdings. Scholars of American Jewry are 
particularly reliant on The Library's archives. The Library is also the 
center for scholarly dialogue within The Jewish Theological Seminary 
(JTS) and the focal point for numerous public events devoted to 
Jewish culture and books. 

Currently, The Library employs 15 librarians, five subject 
specialists and eight support staff. JTS has 40 faculty and 200 staff 
serving 500 students. In fiscal year 2010, after budget reductions, The 
Library's operating budget of $1.7 million plus approximately a half- 



156 A Case Study at The Library of The Jewish Theological Seminary 

million dollars in designated funds each year. Budget designated to 
digital collections is lest than $12,000 per year. 

In The Library's 2005 strategic plan, digitization is emphasized as 
a method of preservation of the original item, because "more damage 
is done to rare materials through human handling than by any other 
means." "The more readers can gain access to images of rare materials 
and forego handling the materials themselves," the document 
continues, "the better they will be preserved." (Strategic Plan, 2005.) 

Large external digitization projects such as The Google Book Project (Google Book, 
2010) and the licensing of the Otzar Ha-Hochma collection of 19,000 fully-digitized 
Hebrew Books and HebrewBooks.org are viewed as efforts to take care of the need 
to digitize modern printed books, allowing the Jewish Theological Seminary library 
to focus resources on only digitizing materials unique to the library's collection. 

Digital Library Strategic Plan 

After successfully completing a number of small digitization projects it 
was clear that a strategic direction for The Library was necessary. In 
November 2005, we participated in the "Persistence of Memory: 
Stewardship of Digital Assets" conference presented by the Northeast 
Document Conservation Center (NEDCC) and partially funded by the 
Institute of Museum and Library Services (IMLS). At that conference, 
Tom Clareson, from PALINET, was soliciting libraries to participate in 
a study funded by the IMLS to assess institutions "digital readiness". 
JTS volunteered and became one of the sample institutions for this 
study. 

Two members of the project team, Tom Clareson (Program 
Director for New Initiatives, PALINET) and Richard Kim (then Digital 
Projects Manager, Metropolitan New York Library Council), served as 
the project evaluators. The assessment entailed completion of a 
detailed survey that posed questions about the existing digitization 
environment, plans, hardware, software, and digital preservation, 
included was a site visit. During the site visit, discussions were held 
with people responsible for selection of items for digitization, creating 
metadata, scanning technicians, conservators, reference librarians 
and administrators and the JTS IT department. The outcome of the 



A Case Study at The Library of The Jewish Theological Seminary 157 

completed survey and the meeting were the basis of a strategic plan 
for The Library's digital program. 

The key interests of the library administration and staff were to 
determine the best way to manage the existing and future digital 
assets; evaluate current status of access to digital materials; to develop 
strategies for future digital projects and funding for those projects; 
and to begin to look at the sustainability of digital projects within The 
Library's digitization program. 

The assessment report outlined the following recommended 
(Clareson and Kim, 2006): 

• Identification and selection of priority collections to be digitized, 
including the development of a written "Digital Collection 
Development Policy." 

• Identification of the primary and secondary audiences for the 
existing and future digital collections of JTS. 

• Exploration of existing standards in content creation and 
metadata. 

• Consideration of upgrades to digitization equipment for future 
projects, as well as exploration of working with digitization vendor 
services. 

• Adoption of a library-wide set of digital procedures. 

• Working with a metadata consultant to make sure "retrospective" 
metadata and the metadata for future projects allows researchers 
to easily reach the JTS digital collections. 

• Exploration of the utilization of a digital asset management system 
to deliver JTS images to its constituents and to the world. 

• Adoption of handling guidelines for materials to be digitized. 

• Documentation of current digital preservation practices, and 
working with a collaborative or third-party digital archiving 
service. 

Implementation of the recommendations began with a document 
outlining potential unique collections to be digitized followed by 



158 A Case Study at The Library of The Jewish Theological Seminary 

guidelines for standards in content creation and metadata and 
establishment of library-wide digitization procedures. Decisions were 
made regarding the upgrade of existing metadata. A digital assets 
management system was licensed and there were explorations 
regarding digital preservation services. 

JTS, as a small academic institution could not support an open 
sources system, such as DSpace (About DSpace, n.d.) or Greenstone 
(Greenstone Digital Library Software, n.d.). Our IT staff is too small 
and turn over was too high in both the IT department and The 
Library. We wanted to host our digital assets locally. We had a long- 
standing relationship with ExLibris, as we run the ALEPH500 system 
at The Library. A decision was made to license DigiTool (ExLibris 
Digitool, n.d.), their digital assets management system. 

One of the significant unique collections identified for digitization 
were the audio-visual collections. These recordings, in multiple 
obsolete formats, are completely inaccessible it is therefore imperative 
for us to digitize. They include musical recording, lectures, films and 
other unique and unidentified materials. 

With recommendations from the digital assessment in hand The 
Library applied, and was awarded, funding from the National 
Endowment for the Humanities (NEH) to assess the audio-visual. 
This assessment conducted by specialists in the field of sound 
preservation, in combination with cross-training of staff with 
responsibilities related to the sound collections, has provided the best 
foundation for conserving primary sound materials and converting 
them, when necessary, and using the best technology to more 
contemporary media. 

The Library created an inventory of the collection, conducted an 
assessment of the current status of the collection with sound 
consultants Chris Lacinak, President of Audiovisual Preservation 
Solutions, and Kevin O'Neill. An assessment report was produced 
and staff was trained to handle the audio-visual collections and 
manage preservation of the collections. (Lacinak and O'Neill, 2008) 

With the recommendations of Audiovisual Preservation Services, 
The Library has moved forward to the next step as laid out in the 



A Case Study at The Library of The Jewish Theological Seminary 159 

assessment report. We are seeking funding to gain physical and 
intellectual control of the sound collections. We have prioritized the 
preservation of certain collections; and have prepared selected items 
for reformatting. Once funding is secured we will be able to perform 
necessary conservation, reformatting, provide for digital preservation 
and make the recordings available on the web. 

One of the most significant issues with which we need to grapple 
is the recommendation of Audiovisual Preservation Services to 
develop in-house capability for the reformatting of our sound 
collections. This will require commitment from The Jewish 
Theological Seminary to build capacity for reformatting on site in the 
existing and underutilized sound studio. The institution would need to 
repurpose the space and commit to allocating funding to acquire 
equipment and staff the lab. The Report states that it would take eight 
years to complete the reformatting of The Library's collection. 

Subsequent to the funded assessment of the sound collection, The 
Library received a large gift of a significant ethnomusicological 
collection and submitted an application for the processing and 
reformatting of these sound collections. 

We received a grant from the Metropolitan New York Library 
Council (METRO) to digitize a collection of 250 wedding poems, 
unique poems composed for wedding parties, particularly in Italy 
during the I7th-i9th centuries. One of our staff members was familiar 
with DigiTool and he was successful in bringing up the system, with 
the Wedding Poems on schedule in January 2008. 

Setting guidelines and standards for the presentation of the 
digital library generated debate amongst the staff. A standing 
committee for creation of the Digital Collections met regularly to set 
guidelines and monitor the development of the digital collections. The 
committee consisted of the Director of Library Services, 
Administrative Librarian for Technical Services, Administrative 
Librarian for Special Collections and the Systems and Digitization 
Librarian. Guidelines were set for a digital file naming convention; for 
the quality of the digital image, and the web interface for the digital 
library. Importantly, decisions were made as to the level of metadata, 



160 A Case Study at The Library of The Jewish Theological Seminary 

how the metadata would be presented (DigiTool presents metadata in 
Dublin core), where the metadata should be created, if none existed, 
and the extent to which existing metadata needed to be upgraded. 

It was determined that The Library would only digitize materials 
that were not under copyright. The exceptions to this are student 
undergraduate senior theses and master's theses where we have 
permission from the students to post their work. 

Funding opportunities 

Digital project comprise of a number of parts, namely, metadata 
creation, digitization of the objects, and bringing the objects up as part 
of the digital library. Each component is labor intensive, and therefore 
costly. One of the major costs that cannot be carried as part of the 
library's budget is the actual digitization of materials. As a result 
external funding and partnerships are required for each special 
project or collection. Funding has been obtained from various sources, 
including public funding, private foundations, non-for-profit 
organizations and individuals. 

Two assessments have been completed and a third in process with 
grants from public sources. The digital readiness assessment was 
funded by an IMLS grant awarded to the NEDCC. The NEH funded 
the assessment of the audio-visual collection. Currently we are 
evaluating the effectiveness of the use of our digital library with funds 
from METRO who receives the funding from New York State. 

Individuals have donated funds to digitize specific collections 
such as the bookplate collection, the Solomon Rosowsky Field 
Recordings fro pre-state Israel (1936 and 1938) and the recordings of 
Cantor Samuel Hallegua from Cochin, India (1967-1968). This was all 
achieved through unsolicited funds. Each of the individuals either had 
a connection to the collection (the Hallegua recordings) or were 
interested in the subject areas (book plates and field recordings). 

Funding has been received in exchange for "in kind" services. The 
Library's entire collection of 35,000 fragments from the Cairo 
Genizah, literally a hidden treasure. These documents are a 
storehouse of documents found in a synagogue Cairo at the end of the 



A Case Study at The Library of The Jewish Theological Seminary 161 

19th century. They were digitized through an international project, the 
Friedberg Genizah Project, where photographers were sent to The 
Library to work on site, but were paid directly by the external project. 

A private foundation had particular interest in digitizing specific 
"treasures" from The Library's collection. This enabled us to digitize a 
number of The Library's most valuable and important illuminated 
Hebrew manuscripts including the Esslingen Mahzor (Ashkenaz, 13th 
century), subsequently loaned to the Metropolitan Museum of Art 
during 2008-2009 and the Rothschild Mahzor, (Florence, 1490). 

Funding for specific projects enable us to select the materials and 
scale the project accordingly. When we began digitization in 2000 we 
selected materials that could be digitized on a flat bed scanner. The 
National Foundation for Jewish Culture awarded funding to digitize 
items related to Jewish American culture. With two grants we 
digitized 350 newspaper clippings from the US press from the 18th 
and 19th centuries and 100 unique pamphlets from the late 18th and 
19th centuries. With funding from METRO in 2008 we digitized 250 
wedding poems. With current funding from METRO we are digitizing 
diaries, in 25 volumes, belonging to the great Jewish theologian 
Mordecai Kaplan. 

Building Digital Projects 

The "NINCH Guide to Good Practice in the Digital Representation 
and Management of Cultural Heritage Materials" (The NINCH Guide 
to Good Practice in the Digital Representation and Management of 
Cultural Heritage Materials, 2002) served as a guide for planning 
specific projects. It emphasizes the many components necessary to 
create a successful digital project. Over the past decade, The Library 
has learned through experience to improve the methodologies, 
workflows, image and digital file formats and minimum metadata 
necessary to create a successful and sustainable for digital projects. 
Each project has served as a building block for improvement. 

A medium-size academic library cannot sustain a full service 
digitization lab with cameras, scanners, audio conversion equipment, 
that quickly become obsolete. The acquisition and maintenance of 



162 A Case Study at The Library of The Jewish Theological Seminary 

technical equipment fell outside our core mission. Our collection 
includes extraordinarily rare materials that cannot leave the premises. 
Items on parchment cannot leave the temperature and humidity 
controlled environment of the rare book room and special reading 
room for extended periods of time. 

We did purchase a flat bed scanner, a digital camera with lights 
and a book scanner. Most of this equipment is out of date but still 
serves us well for in-house work. 

Libraries are experienced in creating metadata for bibliographic 
description of books and other materials. We need to apply those 
methodologies to the digital object and learn how descriptions of 
digital materials differ, or do not, from description of the physical 
object. We have retooled and learned different metadata schemas 
commonly used for digital collections such as Dublin Core (Dublin 
Core Metadata Initiative, 2010) and MODS (Metadata Object 
Description Schema, 2010). 

Prioritizing Projects 

Because of the rarity and research value of the collections, many 
institutions approach JTS with invitations to participate in 
collaborative digital projects. Some funding sources have approached 
the institution about specific digital projects. While prioritizing which 
collections should be digitized funding organizations and individuals 
approach The Library with their projects and request our 
participation. 

The following questions need to be asked before partnering: 
Would a project provide complimentary materials to JTS 
holdings? What collaborations are possible? How important and 
useful are the items to users? How does one assess usefulness? 

The physical condition of the item, the ease and method of 
digitization all play a role in prioritization. Do you digitize the 
"popular" favorites of The Library or do you make little know, hidden 
and inaccessible collections available? When approached by a funding 
group with a specific project how do you evaluate the project in 
relation to overall library priorities? 



A Case Study at The Library of The Jewish Theological Seminary 163 

The Library's goal is to present a broad range of digital materials 
reflecting the breadth of its collection. To date we are presenting 
books, manuscripts, musical scores, sound recordings, archival 
finding aids, flat items, pamphlets, and little know and inaccessible 
collections such as the wedding poems and bookplates. 

Early Projects 

The Library's early digitization projects were flat objects 
(newspaper clippings), pamphlets which were saved as .pdf files 
(American Judaica pamphlets from late 18th and 19th centuries) or 
digitization from slides and transparencies (ketubbah— marriage 
contract - collection). 

Funding was secured for the digitization of the newspaper 
clippings and the pamphlets and for the purchase of a flat bed 
scanner. The files were all saved as .pdfs. Metadata existed in The 
Library's ALEPH500 catalog. The .pdf files were loaded onto the 
ALEPH server and links were made within the bibliographic record 
utilizing the MARC 856 field. 

Access to unique visual materials was a high priority. The 
ketubbah (Jewish marriage contract) collection comprises of more 
than 500 contracts, many of which are illuminated or illustrated. 35 
mm slides or 4 x 6 transparencies were available for about 450 of the 
contracts, all were cataloged. A slide and transparency adapter was 
purchased for the flat bed scanner and they were easily scanned in 
three sizes, thumbnail, medium and large. Each of the three files were 
saved on the ALEPH server and then linked to the bibliographic 
record. Users have access to the small, medium and large sized of each 
of the marriage contracts. 

Partnership Projects 

Partnership projects give us the opportunity to have large 
collections digitized with external funding and to utilize technology 
that is not available in The Library. Crafting a partnership agreement 
has been one of lessons learned over the past five years. One of the 
important issues emerging from these partnership projects, is control 
of the images. The partner may have a copy for use, generally on their 



164 A Case Study at The Library of The Jewish Theological Seminary 

site, but JTS must own the images. Quality of the images is also an 
area of concern with partnership projects. 

We have participated in three major partnership projects each 
with successful outcomes but not without challenges. 

Genizah fragments: A partnership agreement was signed with 
The Friedberg Genizah Project (FGP) to digitize the 35,000 fragments 
from the Cairo Genizah owned by JTS. These documents, discovered 
at the end of the 19th century are "a collection of fragmentary Jewish 
texts stored in the loft of the ancient Ben Ezra Synagogue in Cairo, 
Egypt between the 8th and 19th centuries. These manuscripts outline 
a 1,000-year continuum of Middle-Eastern history and comprise the 
largest and most diverse collection of medieval manuscripts in the 
world." The FGP was established to facilitate and transform Genizah 
research by identifying, cataloging, transcribing, translating the 
fragments, rendering them into digital format (i.e., photographing) 
and publishing them online. (Genizah Project, Executive Summary, 
n.d.) 

FGP hired digital photographers with state of the art digital 
cameras to produce 600 dpi color images of the 35,000 fragments in 
The Library. They funded staff to prepare the material for digitization, 
work with the digital photographers, re-shelve the material and 
document the work. The photographers shot images for about four 
months. All the fragments were shot with targets that included their 
local identification number. The digital images were delivered in 
"dng" format for the repository copy and in .jpg format for access. 
They are available, with registration, on the FGP site. JTS holds an 
internal access copy but has not made it public because of the limited 
metadata accompanying the image. 

Lessons learned from this first partnership project: In order to 
successfully complete a major digitization project one must seize an 
opportunity. This was an outstanding opportunity to digitize the 
35,000 Genizah fragment. The project was driven by the funder. They 
determined the items to be digitized, the method of digitization, the 
choice of photographer, the rate of digitization, etc. There was no 
specific budget for the completion of this work nor was there a 



A Case Study at The Library of The Jewish Theological Seminary 165 

designated time-frame. Work commenced and continued at a swift 
pace until it was completed. The funder's representative wanted to be 
involved in The Library's internal workflow. Often work was rushed 
and not adequately planned. The most significant issue is that there is 
no metadata for the Genizah fragments and the material is not freely 
available. Only scholars who register, and are accepted by, the FGP 
may search the information. 

Library "treasures": There was another funder who selected 
items which he was interested in digitizing, supplied the photographer 
(who was approved by The Library) and delivered digital images in 
TIFF format. While this project was funder driven, it supplied The 
Library with images of some of its greatest manuscripts such as the 
Rothschild Mahzor (MS 8892) from Florence, 1490. A separate site, 
Special Treasures from The Library of The Jewish Theological 
Seminary (http://jtslibrarytreasures.org/) with turning the page 
technology was launched. Only 50 of the 450 leaves are on display on 
this site. The site also includes other digitized manuscripts from the 
collection, which are available with full metadata, on The Library's 
digital library in 2008. 

Rare book digitization: A project to digitize very rare books, 
incunabula and 16th century books, was completed during the fall of 
2009. This project reaped the benefits of our experience. As this was 
not a comprehensive project, books needed to be selected, posing a 
more complicated workflow. The goal was to digitize only unique 
books that had not been digitized by other libraries. Once they were 
identified, the condition of the book was checked with guidelines from 
the paper conservators. The books were digitized on a Zeutschel 
OS12000 book scanner in black and white. Each book needed to have 
a wide enough opening so that the image could be shot. A ruler was 
placed on each page. All the books were cataloged and the file name 
for the digital image was the ALEPH (our ILS) system number for the 
book. The bibliographic records were retrieved from ALEPH. 
Hebrewbooks.org, the organization that was performing the 
digitization, delivered the digital images in tiff and pdf formats. We 
loaded the pdf files to our server and are now in the process of loading 
them to digital library. We exported the metadata and delivered it to 



166 A Case Study at The Library of The Jewish Theological Seminary 

Hebrewbooks.org as they are loading the pdf files and metadata to 
their collection of full text religious books in Hebrew which to data 
total more than 40,000 titles. 

Lessons learned: We took far more control over the steps of this 
project. The external funder did however chose the equipment used, 
we would have preferred to digitize the books in color, utilizing a 
digital camera. At the beginning of the project the digitization 
technician was trained by The Library's paper conservator in order to 
ensure careful handling of each book. Greater control over the choice 
of materials, the way the materials were handled and preparation of 
the metadata made this a successful project. During November 2009 
92 rare books were digitized. 

Outsourcing 

We have no capacity to digitize sound at JTS. We have digitized 
two musical collections working with two external sources. One 
collection was digitized by a private sound engineer and the other 
project went to the lab at the Center for Jewish History, a non-for- 
profit organization that had until our request, not digitized for other 
institutions. This proved to be beneficial for them as it provided them 
with some income during a "down" period in their digitization lab. 
Plans are underway to send an archival collection, comprising 
volumes of diaries, to the studio of a digital photographer that we have 
worked with on numerous occasions. 

The output of the digitization of the 252 Solomon Rosowsky Field 
Recordings collection was in .mp3 format. We requested an archival 
output. This resulted in a raw and sometime difficult to hear output. 
The metatdata for this collection, created in an Excel spreadsheet, was 
created at the time of digitization and is not part of the catalog as they 
are a subsection of an archival collection. 

The quality of the sound the Hallegua collection, digitized in 
2009, proved to be far more successful. Partnering with the Center for 
Jewish History and utilizing their digital lab we accepted their 
recommendations and created archival .wav copies and access .mp3 
copies. The quality of the work completed in the second project is 



A Case Study at The Library of The Jewish Theological Seminary 167 

clearer than the first project. Metadata is yet to be created for this 
collection. Once it is available we will make the collection accessible. 

Library initiated project: 

When we have the luxury of planning and implementing a digital 
project on our own terms we have learned that this is the ideal 
situation for us. The digitization of manuscripts from a pre-selected 
list has been the most successful project, from a digitization point of 
view, that we have embarked on. We selected important manuscripts 
that had no surrogates. The processes of the project were reviewed by 
a project team who included: paper conservators, conservation 
assistant, curator, special collections reading room staff, systems and 
digitization librarian, heads of technical services and special 
collections and director of library services as chair. The goal was to 
look at the entire scenario of bringing in a digital photographer to 
shoot 4,000 manuscript leaves over a two-week period. We planned 
the space for digitization, the spreadsheet of targets, the physical 
targets, workflow for pulling items and reshelving them, assistance for 
the digital photographer, metadata for the manuscripts, and plan for 
bringing up the digital objects in the digital library. 

The planning meeting raised questions from all the stakeholders. 
If there were any outstanding questions, all stakeholders were later 
updated on the answer or resolution. Meeting with the entire group 
enable everyone involved to know exactly who had a stake in the 
project. It allayed any fears and anxieties. Bringing the team together 
enabled a smooth process for the preparation of the materials for 
digitization, the actual digitization and the preparation of the 
metadata. The final piece, bringing the digital objects up in the digital 
library posed other problems regarding staffing. 

Challenges 

Much has been accomplished but not without challenges. Hardware 
and software capacity, staffing and funding are all issues that we 
continue to grapple with. 



168 A Case Study at The Library of The Jewish Theological Seminary 

Capacity issues 

The capacity of The Library's server that stores digital objects and 
the capacity of the library's license for its DigiTool installation are 
both limited. Issues related to capacity entail actual storage of the 
digital images. Often one has different manifestations for a single 
image or multiple images for a single document - for example, for a 
five-page document that was shot as a RAW file there are TIFF 
images, which in turn may generate JPEG files and also a PDF. 

The number of version of the same document that one needs to 
keep is determined by capacity. There is the repository server where 
one stores the images. The number of manifestations that are saved 
there may be determined by the institution's capacity. One must work 
with those responsible for the institution's servers. 

Staffing 

Staffing is key to the success of each digital project. The key to 
building a successful team for the digital library is for each team 
member to be cognizant of their specific role in the project. 

The team consists of curators who select materials for 
digitization. The selections are determined by funding, importance, 
fragility of the item, value of material, or by other criteria. 

Once the items are selected they need to be assessed by the 
paper conservators (for items on paper and parchment). They look 
at the physical stability of the item. They can recommend if an item 
can be digitized, with recommended equipment (scanner, book 
scanner, photographer). They can also recommend if an item needs to 
be encapsulated in Mylar before digitization, and/ or if an item needs 
to have a conservation assistant working with the photographer 
throughout the shoot. 

Special collections staff is responsible for pulling and shelving 
books. They need a complete and accurate list of the items to be 
scanned. For a longer-term project, they need to prepare materials 
each day to be digitized and materials that need to be reshelved. If the 
project is based on a list of items found in a certain collection, they 
may often need to search for the item. 



A Case Study at The Library of The Jewish Theological Seminary 169 

Digital photographer or scanner. Generally, this person is 
not part of The Library's staff, but is an outside contractor or a 
temporary worker hired to do the specific work. They need to be 
integrated into the working environment of The Library, which may 
include the physical location of the equipment, the rules and hours of 
the special collection reading room. 

Technical assistant to assist the photographer with turning 
pages, preparing the shoot, and other needs surrounding the shoot. 

Cataloger and metadata librarian needs to review the 
current metadata for the object and make appropriate adjustment to 
the metadata for the digital library based on the standards 
established. Some of the metadata added include genre, collection, 
and location of digital object. In some cases the metadata needs to be 
upgraded, but this is subject to great debate amongst library staff. In 
most cases it depends on the amount of existing metatdata and the 
object in hand. 

Systems and digitization librarian - Needs to export 
metadata and deliver it to the cataloger and metadata librarian and 
needs to prepare the materials for ingest into the digital library. 

Since the existence of the JTS digital library, in January 2008, we 
have had three systems and digitization librarians. Because of the staff 
turnover, it has been challenging to complete many of our digital 
projects and make them available online. Since October 2009, with a 
new person in this position, new workflows are being documented and 
the actual ingest of digital objects is being completed by other 
members of the staff under the guidance of the systems and 
digitization librarian, rather than by the person holding that position. 
The position has evolved, with the initial incumbent successfully 
building and launching the digital library. The digital library was 
launched and there were initial documents added to the collection, 
there was no documentation of the processes used which left the 
second person in a difficult situation. There was little activity in 
loading objects during that year. With the third person hired in the fall 
of 2009 it was clear that the first step was to create workflows and to 
clearly document them for others to follow and for The Library to have 



170 A Case Study at The Library of The Jewish Theological Seminary 

in order to successfully maintain the digital library. Lessons learned: 
stop and document what you are doing and share the documentation 
and the work. 

From Digital Projects to a Digital Program 

As we move into the second decade of digitization these are some of 
the strategic directions that we need to address. 

Assessment: With funding from METRO, The Library will assess 
the effectiveness of its digital collections in DigiTool. A survey is 
currently being built and will be posted on the site and sent out to a 
wide group of constituents. The plan is to implement 
recommendations from the survey, add new collections to the site, 
most notably the Mordecai Kaplan diaries, and then survey the 
effectiveness at the end of the grant period. We hope that this 
assessment will help guide us in our work. 

Documentation: We are documenting the workflows that we are 
establishing for the creation of digital collections. With the 
documentation, and with adequate funding, we will be able to hire 
staff to work on building the digital collections. 

Digital preservation: In the winter of 2008 we participated in the 
Sustaining Digital Preservation workshop funded by the IMLS. We 
also attended a METRO meeting in the fall of 2008 focused on digital 
preservation. Conversations have begun with ExLibris about their new 
digital preservation software. We need first and foremost to obtain the 
commitment of the JTS administration to invest in digital 
preservation. This is challenging, considering the serious nature of the 
institution's finances in wake of the economic downturn in the US. 
Little work has been done at JTS to address these needs. 

New projects, improving workflow, increasing capacity, 
increasing staff to enrich with metadata are all necessary aspects of a 
successful digitization program. Our hope is that with creativity, 
imagination, thoughtfulness and good luck, organizations drawn to 
the uniqueness and richness of our collections will continue funding 
The Library of The Jewish Theological Seminary in this important 
work. 



A Case Study at The Library of The Jewish Theological Seminary 171 



References 

About DSpace. (n.d.). Retrieved March 9, 2010 from 
http : / / www. dspace.org/ about-dspace/introducing/ 

Clareson,Tom and Kim, Richard. (2006). Digital Program Survey 
Final Report from April 5, 2006 Visit. 

Dublin Core metadata initiative. (2010). Retrieved March 30, 2010, 
from http:/ / dublincore.org 

ExLibris DigiTool. (n.d.) Retrieved March 9, 2010 from 

http:/ /www.exlibrisgroup.com/ category/DigiToolOverview 

Genizah project, executive summary. ( n.d.) Retrieved March 30, 
2010, from http://www.genizah.org/about- 
Executive_Summary.htm. 

Google Books. (2010). Retrieved March 13, 2010 from 
http://books.google.com/ 

Greenstone digital library software, (n.d.). Retrieved March 9, 2010 
from http://www.greenstone.org/ 

Lacinak, Chris and O'Neill, Kevin. (2008). Sound Recording 
Consultation Final Report, June 26, 2008 

Metadata object description schema. (2010). Retrieved December 11, 
2009, from the official web site 
http : / / www.loc.gov/ standards / mods / 

Strategic Plan. (2005). The Library of The Jewish Theological 
Seminary. 2005. 

The NINCH guide to good practice in the digital representation and 
management of cultural heritage materials. (2002). Retrieved 
May 9, 2010 from 

http://www.nyu.edu/its/humanities/ninchguide/ 



Local Objects, Local People, Local History: 
Creating the Wisconsin Decorative Arts 
Database 

Emily Pfotenhauer (Wisconsin Heritage Online, Wisconsin 
Historical Society and the Chipstone Foundation) 



Abstract 

The Wisconsin Decorative Arts Database is a digital collection of 
three-dimensional artifacts from the collections of historical societies 
and museums throughout Wisconsin, hosted by the Wisconsin 
Historical Society and supported by the Chipstone Foundation. Since 
2006, the project has documented nearly 1,000 examples of furniture, 
ceramics, textiles and other decorative arts made by early Wisconsin 
craftspeople and held in the collections of over 40 institutions 
throughout the state. This case study examines the genesis of the 
project, the photographic standards and metadata specifications 
established for object documentation, and the unique challenges of 
developing a diverse digital collection of museum artifacts from a wide 
variety of local and regional institutions. 

Keywords: Artifacts, Collaboration, Historical societies, Local 
history, Metadata, Museums, Photography 



In the past decade, the work of state- and regionally-based digitization 
programs across the country has resulted in an explosion of local 
history resources available online. These initiatives are collaborative 
efforts among libraries, archives, and museums to make their 



172 



Local Objects, Local People, Local History 173 



collections freely available to a broad audience of students, teachers, 
historians, and genealogists. Some of the earliest and most influential 
of these programs include the Maine Historical Society's Maine 
Memory Network, the Minnesota Digital Library, and the Colorado 
Digitization Project (now known as the Collaborative Digitization 
Program at the Bibliographic Center for Research). For the most part, 
programs such as these have focused on what some museum 
professionals refer to as "the flat stuff: photographs, books, 
manuscripts, and other two-dimensional materials. This focus on the 
digitization of archival materials belies the fact that three-dimensional 
artifacts form the core of the collections of most local historical 
societies and museums. Moreover, the everyday objects people create 
and consume (for example, furniture, clothing, and tools) are 
considered by many scholars to offer significant evidence for historic 
research, as much if not more so than historic photographs or original 
manuscripts. Yet most institutions participating in collaborative 
digitization programs have not made their collections of three- 
dimensional objects available online at the same rate as their two- 
dimensional holdings. 

Case studies of statewide and national digitization programs 
reveal the challenges common to most collaborative digital endeavors, 
regardless of the types of material being digitized. Roberto (2008) 
describes the lack of consistent standards in place for describing 
collections among even the largest and best-staffed museums in Great 
Britain. When collaborations are expanded to include small, local 
historical societies and museums, the creation of consistent 
collections data becomes even more challenging due to a lack of 
available staff, resources, and expertise (Rowe & Barnicoat, 2009). 
These obstacles are compounded when working with three- 
dimensional objects, which require more space, equipment, and 
technical expertise to photograph and more research and subject 
expertise to catalog. 

This chapter examines the development of one digital collection 
of three-dimensional objects from multiple institutions: the Wisconsin 
Decorative Arts Database (http://content.wisconsinhistory.org/ 
decorativearts). This project was initiated in 2006 by the Wisconsin 



174 Local Objects, Local People, Local History 

Historical Society and the Chipstone Foundation of Milwaukee, a 
private foundation for the study of American decorative arts and 
material culture. Chipstone and the Society had three goals in mind at 
the outset of the project: l) to bring to light objects held in the 
collections of Wisconsin historical societies and museums and share 
them with a broader audience, 2) to document examples of furniture, 
ceramics, textiles, and other decorative arts made by nineteenth and 
early twentieth-century Wisconsin craftspeople in order to reveal 
settlement patterns and the persistence of handicraft traditions in the 
state, and 3) to add the first three-dimensional artifacts to Wisconsin 
Heritage Online (http://www.wisconsinheritage.org), a research 
portal that brings together a diverse range of digital collections from 
Wisconsin libraries, archives, museums, and historic sites. 

The Wisconsin Decorative Arts Database is founded on the work 
of a multitude of "real world" individuals— the staff and volunteers at 
local historical societies and small museums throughout Wisconsin 
who work to preserve and share their collections, as well as the men 
and women who built the chairs, stitched the quilts, and threw the 
pots that are recorded in the database. In addition, the story of the 
Wisconsin Decorative Arts Database is a highly personal "real world" 
story for me, as the project grew out of my interest in the material 
culture of my home state. 

The seeds were planted in 2005 while I was a graduate student 
studying American furniture and decorative arts in the Department of 
Art History at the University of Wisconsin-Madison. The time had 
come for me to choose a subject for my master's thesis research, and I 
wanted the opportunity to study objects hands-on and up close. I 
decided to focus on furniture made in the community of Mineral 
Point, an important early trade center in southwest Wisconsin— as 
well as the town where I grew up. During the year I spent researching 
and writing my thesis, I unearthed a surprising amount of locally- 
made furniture still surviving in both public and private collections 
and met numerous Mineral Point residents dedicated to preserving 
their community's distinctive history. 



Local Objects, Local People, Local History 175 



This search for furniture made in Mineral Point caught the 
attention of the Chipstone Foundation of Milwaukee. In 2006, 
Chipstone approached the Wisconsin Historical Society with the idea 
to create an expanded fieldwork program that would include 
furniture, ceramics, textiles, and metalwork made by craftspeople 
throughout Wisconsin between 1820 and 1920. From the beginning, a 
central focus of both Chipstone and the Society was to make these 
findings freely available to the public in digital form. Both institutions 
were already leaders in the digitization of cultural heritage materials 
— Chipstone with its Digital Library for the Decorative Arts and 
Material Culture (http://digital.library.wisc. edu/i7ii.dl/DLDecArts), 
developed in collaboration with the University of Wisconsin Digital 
Collections Center, and the Society with a number of innovative 
digitization efforts including the award-winning American Journeys 
(http://www.americanjourneys.org) and Turning Points in Wisconsin 
History (http://www. wisconsinhistory.org/turningpoints)— so a 
digital database of museum collections was a natural extension of both 
institution's missions. 

I was appointed as Chipstone's Charles Hummel Fellow to 
manage all aspects of the initiative— fieldwork, photography, research 
and writing as well as the structure and organization of the digital 
collection. As my background was in art history, not digital 
collections, I had a great deal to tackle in order to create an effective 
online resource. What kinds of materials would be included and how 
would they be located? How would each object be imaged? How would 
users search the database? What kinds of metadata should be 
included? And what exactly was metadata, anyway? 

Fortunately, a few pieces of the puzzle were already in place. The 
Wisconsin Historical Society would host the collection in 
CONTENTdm, a content management system already in use by the 
Society's Library-Archives division. Wisconsin Heritage Online 
(WHO), a developing statewide digitization program, would also 
harvest the collection into its central web portal. WHO's Metadata 
Guidelines (2006) provided a useful introduction to creating and 
organizing digital content, as did the Collaborative Digitization 
Program's Dublin Core Metadata Best Practices (2006). However, 



176 Local Objects, Local People, Local History 

these guidelines were established specifically for images and text- 
based materials, not three-dimensional artifacts. 

Translating information about chairs, quilts, and pots into the 
standardized fields recommended by existing guidelines was a bit like 
trying to fit square pegs into round holes. Developing a crosswalk or a 
systematized data import process was impossible, because no two 
museums or local historical societies handle cataloging in the same 
way. In fact, many institutions do not even use a digital collection 
management system, instead relying on typed index cards or 
handwritten ledgers to catalog their collections. 

I developed a standard set of metadata elements based on WHO 
recommendations, mapped them to the fields defined by the Dublin 
Core (http://dublincore.org/documents/dcmi-terms), and assigned 
customized local field names. For example, the Dublin Core Creator 
element became a local "Maker" field, while Dublin Core's 
Format.medium became a field called "Materials and Techniques." In 
addition to providing images and physical descriptions of each 
documented object, a central focus of the project was to place each 
artifact in its historical context by conducting research on the 
biographies of makers and owners as well as design influences and 
cultural meanings. This goal called for the creation of multiple Dublin 
Core Description fields: one for physical description, one for historical 
information, and one for the citation of research sources. 

Another key cataloging decision was the selection of a controlled 
vocabulary appropriate to the database content. Because it is built into 
PastPerfect Museum Software, a popular content management 
system, The Revised Nomenclature for Museum Cataloging, aka 
Chenhall's Nomenclature (Blackaby, Greeno, and the Nomenclature 
Committee, 1995) is used by many local historical societies and small 
museums in Wisconsin. However, the Getty Research Institute's Art 
and Architecture Thesaurus (http://www.getty/edu/research/ 
conducting_research/vocabularies/aat) provides more nuanced and 
specific terms for decorative arts objects. A significant level of 
granularity is required to accurately assign object names and subject 
headings in the database. Using Chenhall's Nomenclature, a chair can 



Local Objects, Local People, Local History 177 



be described primarily according to function, e.g. "Chair, Dining" or 
"Chair, Side." The AAT allows for more precise descriptions of form 
and style. For example, a chair can be distinguished as a bow-back 
Windsor chair, a fan-back Windsor chair, or a sack-back Windsor 
chair. While the differences may appear minute to a general audience, 
these kinds of distinctions are essential to the work of decorative arts 
scholars and collectors. — A revised and updated edition of Chenhall's 
Nomenclature, Nomenclature 3.0, offering a greatly expanded 
vocabulary for the description of museum artifacts, was released in 
late 2009. Future subject headings provided in Wisconsin Decorative 
Arts Database metadata may incorporate terms from this new edition. 

One major question remained before any fieldwork could begin: 
how to create quality digital images of objects that not only varied 
widely in size and format but were scattered across the state? Just as 
few institutions could offer detailed digital catalog records for their 
collections, few had the resources or expertise available to image their 
collections extensively, if at all. In addition, as with its metadata 
guidelines, the WHO Digital Imaging Guidelines (2006) addressed 
the scanning of two-dimensional materials almost exclusively and 
offered little support for photographing three-dimensional objects. 
Much of the available documentation on imaging museum artifacts, 
such as Stanford University's Digital Michelangelo Project (Levoy & 
Garcia-Molina, 2000) and the work of the Graphics Lab at the 
University of Southern California's Institute for Creative Technologies 
(Hawkins, Cohen, & Debevec, 2001) centered on the creation of 
complex three-dimensional digital renderings using multiple cameras 
and laser scanning. 

It was necessary to develop a simple, portable imaging approach 
so that photography could be completed quickly (to adapt to the 
limited schedules of volunteers and busy staff) and unobtrusively (to 
accommodate the often cramped storage quarters of small 
institutions). I arrived at my photographic process through a 
combination of consultation with professional photographers and on- 
site trial and error. Jim Wildeman, a photographer based in Madison, 
provided a basic introduction to artifact photography. Selecting the 
right equipment for the job was essential. The Chipstone Foundation 



178 Local Objects, Local People, Local History 

funded the purchase of a Canon Rebel XTi digital SLR camera as well 
as a tripod, a remote shutter release, a neutral gray paper backdrop 
and backdrop stand, and two stationary tungsten lights on adjustable 
stands. Indispensable incidentals included a long extension cord, tape 
measures (both fabric and metal), white cotton gloves for handling 
items, and cloths and brushes for removing dust from artifacts. The 
entire photography kit could be packed into the trunk of a car and 
quickly set up on location. Participating institutions need only to 
provide an electrical outlet and suitable working space. 

After the framework of the digital collection was laid out and 
procedures for photographing and cataloging artifacts were 
determined, on-site work could begin. Based on my experience with 
local historic sites and individual collectors in Mineral Point, I knew 
that putting out a general call for material via an email listserv would 
yield minimal results. It was much more effective to contact each 
potential contributor directly, usually with an introductory email or 
letter and then a follow-up phone call to discuss possible objects 
before scheduling an in-person visit. Once on site, objects were 
selected for documentation based on the availability of documented 
provenance, oral history associated with the item, or stylistic 
similarities to other known objects. 

I began my fieldwork with two familiar sites from my graduate 
research: the Mineral Point Historical Society and Pendarvis, the 
Wisconsin state historic site in Mineral Point. Starting with these sites 
gave me the opportunity to refine my documentation techniques in a 
comfortable environment. I could take as long as I needed to make 
minute adjustments in lighting or re-measure the dimensions of a 
chair or flowerpot. I connected with the next three participants, all 
mid-sized regional museums, via word-of-mouth recommendations: 
the Neville Public Museum of Brown County in Green Bay, the 
Chippewa Valley Museum in Eau Claire, and the Sheboygan County 
Historical Society. 

After working with this handful of pilot institutions, it was clear 
that the phrase "decorative arts," while a common academic term, was 
something of a stumbling block for potential participants. Some 



Local Objects, Local People, Local History 179 



possible contributors read the phrase too narrowly, expressing 
concern that the objects in their collections were not "decorative" 
enough, while others interpreted it too broadly, hoping to include 
items manufactured outside of Wisconsin but used locally or objects 
that lacked any known history whatsoever. At the same time, I was 
discovering many artifacts that did not fit my original parameters but 
were of compelling historical significance. 

Although the title of the database remained the same, the 
selection criteria were adapted to accommodate objects that offered 
important evidence of Wisconsin craft practices and industrial 
development, even if they did not fit into an academic definition of the 
decorative arts. For example, the staff of the Historic Blooming Grove 
Historical Society in Madison contacted me about their collection of 
materials from the Frank J. Hess and Sons Cooperage, one of the last 
manufactories in the nation to produce hand-hewn white oak beer 
barrels. While not "decorative arts" in the traditional sense of the 
term, the Hess barrels reveal a fascinating story of an immigrant 
craftsman who brought a traditional European craft practice to 
Wisconsin, adapted it to the local environment, and passed it on to a 
second generation (Holland, 1966). 

Although I received a number of word-of-mouth 
recommendations and direct contacts, most database participants 
were located through the Wisconsin Historical Society's Directory of 
Wisconsin Local History Organizations (http://www. 
wisconsinhistory.org/localhistory/directory). The Society's local 
affiliates include more than 350 county historical societies, local 
historical societies, community museums and historic house 
museums. The vast majority of these institutions are run by volunteers 
or a tiny cohort of part-time staff. Many are open to the public 
seasonally and for limited hours, sometimes as little as one or two 
weekends each month. This meant that arranging access to collections 
hinged on flexible scheduling and the willingness of staff and 
volunteers to open their doors during non-public hours. 

The fieldwork process, an in-person visit to select, photograph, 
and document individual objects, is the most rewarding, yet also most 



180 Local Objects, Local People, Local History 

exhausting, part of the project. From September 2006 through June 
2009, 1 logged over 10,000 miles traveling to sites across the state. I 
sifted through all manner of collection storage facilities, some 
meticulously organized and others packed to the gills with artifacts 
piled on shelves and stuffed in boxes. I handled thousands of artifacts 
and toured all kinds of historic buildings. Most excitingly, I met 
dozens of people who were passionate about their work, dedicated to 
preserving the history of their communities, and eager to share their 
collections with new audiences. 

Depending on the number of items selected for documentation, a 
site visit can last from a few hours up to two or three full days. 
Preparing the material to post online takes much longer. A series of 
graduate and undergraduate interns from the Material Culture 
Program and the School of Library and Information Studies at the 
University of Wisconsin-Madison have helped support this phase of 
the project. Basic editing procedures, including cropping, sharpening, 
and color balancing, are performed on each image using Adobe 
Photoshop or ACDSee. Images and corresponding metadata are 
uploaded directly to the Wisconsin Historical Society's CONTENTdm 
server. 

A central mission of the database is to establish an historic 
narrative for each artifact by investigating the history of the object's 
maker and owner as well as its cultural significance and formal 
influences. Due to the myriad of primary source documents and other 
historic material now available online, a major portion of this research 
has been conducted solely in the digital realm. The most useful 
sources for biographical and genealogical information proved to be 
the Wisconsin Historical Society's Genealogy Index (http:// 
www.wisconsinhistory.org/vitalrecords), offering pre-1907 birth, 
death, and marriage records for the state of Wisconsin, and the full- 
text-searchable Wisconsin County Histories (http://www. 
wisconsinhistory.org/ wch). The commercial website Ancestry.com 
was also valuable for searching federal census records. 

Participating institutions are required to sign a Memorandum of 
Understanding that authorizes the Wisconsin Decorative Arts 



Local Objects, Local People, Local History 181 



Database to photograph objects and publish those photographs and 
associated metadata online. Each institution receives an archival- 
quality CD containing the image files and retains full rights to those 
images. (A second CD, along with any paperwork generated during the 
research process, is archived at the Wisconsin Historical Society and 
the digital images are backed up on the Society's servers.) For the 
most part, participants have readily agreed to these terms. However, a 
few have expressed concern about making material from their 
collections freely available online, citing either a desire to retain 
control over digital images or a fear of theft of the physical collection. 
In most cases, these fears were abated with the argument that the 
benefits of participation— particularly the increased opportunity for 
exposure and publicity— outweigh the risks. 

Participation offers historical societies and museums the 
important benefit of an increased online presence. Most participating 
institutions maintain their own websites, but many lack the resources 
to fill these sites with substantive information about their collections. 
To increase online exposure, each contributor is clearly identified at 
several points in the database. The metadata for each catalog entry 
includes an "Owner" field as well as a rights statement with a link to 
the institution's own website. The opening page offers users the option 
to browse the database by selecting a specific participant's collection. 
Another page functions as a directory of content contributors, with an 
image and description of each institution as well as a link to their own 
website. 

Two popular free web features— a blog hosted by WordPress and a 
gallery on the photo-sharing site Flickr— supplement the database and 
extend its presence online. I use the blog, Wisconsin Object 
(http://wisconsinobject.wordpress.com), to chronicle my travels and 
announce when new content has been uploaded to the database. It 
also provides a venue for more lengthy examinations of selected 
artifacts and makers. The Wisconsin Decorative Arts Flickr gallery 
(http://www.flickr.com/photos/wisconsindecarts), which is linked 
from the blog, presents examples of recently uploaded database 
content. While not numerous, user comments on both the blog and 
the image gallery have provided an important window into the 



182 Local Objects, Local People, Local History 

audiences for this material and the types of objects they find most 
compelling. For example, the most-viewed image on Flickr is a 
colorful beaded bandolier bag made by Great Lakes Indians in the late 
nineteenth century, now in the collection of the Wisconsin Historical 
Museum. A corresponding blog entry on the cultural significance of 
bandolier bags is one of the most popular posts. Users who have 
commented on the blog or added the image to their own Flickr 
galleries include craft hobbyists and collectors— two major audiences 
that were not considered at the outset of the project. 

The Wisconsin Decorative Arts Database's regional approach to 
material culture research is nothing new. Since the 1970s, researchers 
throughout the United States have worked to document decorative 
arts made locally and held in local collections. Most notably, the 
Museum of Early Southern Decorative Arts (MESDA) in Winston- 
Salem, North Carolina established a fieldwork program in the 1970s 
and 1980s, supported by the National Endowment for the 
Humanities, to locate artifacts made by craftspeople working in seven 
southern states before 1820 (Niven, 2001). This research yielded a 
vast paper-based archive of photographs and files that now serves as 
an indispensable resource for scholars studying the material culture of 
the American South. 

While modeled on the work of MESDA and other statewide 
fieldwork initiatives, the Wisconsin Decorative Arts Database is the 
first program of its kind to be created entirely in digital form. An 
online database of regional artifacts offers a number of advantages 
over a paper archive or a published catalog. New content can be added 
continually, revisions can be made in response to new research, and 
related materials can be brought together via hyperlinks. Moreover, 
making collections available digitally helps small, local cultural 
heritage institutions remain relevant to students, historians, and other 
audiences whose research is grounded in the ready availability of 
online resources. 

Since 2006, nearly 1,000 catalog entries have been added to the 
Wisconsin Decorative Arts Database, representing artifacts from the 
collections of 40 historical societies and museums throughout 



Local Objects, Local People, Local History 183 

Wisconsin. In September 2009, the database was recognized with an 
Award of Merit from the American Association for State and Local 
History. With financial support committed by the Chipstone 
Foundation and the Kaufman Americana Foundation through June 
2011, the program will continue to unearth important evidence of 
early craft production in Wisconsin. In addition to adding new 
content, explorations are underway to develop creative ways to 
expand and promote the project, including both online and bricks- 
and-mortar exhibitions and publications. 

References 

Blackaby, J. R., Greeno, P., &The Nomenclature Committee. (1995). 
The revised nomenclature for museum cataloging: A revised and 
expanded version of Robert G. Chenhall's system for classifying 
man-made objects. Walnut Creek, CA: AltaMira Press. 

Bourcier, P., Rogers, R., &The Nomenclature Committee. (2009). 
Nomenclature 3.0 for museum cataloging: Third edition of Robert 
G. Chenhall's system for classifying man-made objects. Walnut 
Creek, CA: AltaMira Press. 

Collaborative Digitization Program Metadata Working Group. (2006). 
Dublin Core metadata best practices (version 2.1.1). Retrieved 
from http : / / www.bcr.org/ dps/ cdp 

Hawkins, T., Cohen, J., & Debevec, P. (2001). A photometric approach 
to digitizing cultural artifacts. In Proceedings of the 2001 
Conference on Virtual Reality, Archeology, and Cultural Heritage, 
333-342. New York: Association for Computing Machinery, doi: 
10.1145/584993.585053 

Holland, R. (1966). The last wooden beer barrels. Wisconsin Tales 
and Trails 7, 12-16. 

Levoy, M., & Garcia-Molina, H. (2000). Creating digital archives of 
3D artworks [White paper submitted to the National Science 
Foundation's Digital Libraries Initiative]. Retrieved from 
http : / / graphics . stanf or d. edu /pr oj ects/dli/ white-pap er/ dli .html 



184 Local Objects, Local People, Local History 

Miller, S. J. (Ed.)- (2006). Wisconsin Heritage Online metadata 
guidelines (version 2.0). Retrieved from 
https://wiheritage.pbworks.com 

Niven, P. (2001). Frank L. Horton and the roads to MESDA. Journal 
of Early Southern Decorative Arts, 27, 1-147. 

Roberto, F. (2008). Exploring museum collections on-line: The 

quantitative method. In J. Trant and D. Bearman (Eds.), Museums 
and the Web 2008: Proceedings. Toronto: Archives & Museum 
Informatics. Retrieved from http://www.archimuse.com/ 
mw20o8/papers/roberto/roberto.html 

Rowe, P., & Barnicoat, W. (2009). NZMuseums: Showcasing the 
collections of all New Zealand museums. In J. Trant and D. 
Bearman (Eds.), Museums and the Web 2009: Proceedings. 
Toronto: Archives & Museum Informatics. Retrieved from 
http : / / www.archimuse.com/mw2009/ papers / rowe / rowe.html 



Digitizing the Rare Book Collection of the 
Leo Baeck Institute 



Renate Evers (Leo Baeck Institute, New York) 



Abstract 

The article focuses on typical problems encountered during the 
establishment of a digitization program for rare books at a small 
special library - from finding the best solutions for rare book cradles 
to organizational and technical challenges and pitfalls. 

Keywords: Digitization, Judaica Collection, Rare Book Collection, 
Special Library 



The Library of the Leo Baeck Institute: 
A Medium-sized collection 

The Library of the Leo Baeck Institute (LBI) in New York is 
internationally recognized as the most comprehensive repository for 
books documenting the history and culture of German-speaking 
Jewry. Over 80,000 volumes and 900 periodical titles provide 
important primary and secondary material. Most of the collection 
deals with central European Jewry during the nineteenth and early 
twentieth centuries. It also includes material dating back as far as the 
16th century and is as current as the Jewish population in Germany 
today. The focus of our collection is on the diverse culture of German- 
speaking Jewry, especially in the arts, sciences, literature, philosophy, 
and religion. The Institute was established in New York in 1955, one of 
the largest centers of the exiled German-Jewish community. In 2000 



185 



186 Digitizing the Rare Book Collection of the Leo Baeck Institute 



the Leo Baeck Institute became part of the Center for Jewish History, 
an umbrella organization for 5 Jewish research institutions. 

The Rare Book Collection of the Leo 
Baeck Institute 

The Rare Book Collection of the Leo Baeck Institute (LBI) consists of 
ca. 2,000 volumes primarily in the field of German Judaica, dating 
from the earliest period of printing in the 15th century and continuing 
through the annihilation of European Jewry under the Third Reich. 
Rich in rarities ranging from early 16th century writings to Moses 
Mendelssohn and Heinrich Heine, first editions and dedication copies 
of works by more recent prominent writers, many of its volumes were 
salvaged from famous Jewish libraries that were confiscated and 
dispersed by the Nazis. In many cases these primary sources cannot 
be found anywhere else in the world. An important part of the rare 
book collection is a collection of illustrated and art books. Among 
them are many limited editions of twentieth century artists' portfolios 
and several illustrated eighteenth century books on Jewish customs. 
The Institute also has a significant collection of books dealing with 
"Science of Judaism" (Wissenschaft des Judentums). The emergence 
and establishment of modern Jewish scholarship in the 19th and 20th 
century in Germany and Europe became the cornerstone and 
predecessor of Jewish Studies. 

Project: Digitizing the Rare Book Collection of the 
Leo Baeck Institute: 

The goal is to digitize the rare book collection in order to make these 
exceptional primary sources easily available to researchers on and 
offsite, and to minimize use of the originals. To begin, the focus is on 
titles which are in highest demand. 

We started a pilot project with a selection of 38 books in 2005 
when the Leo Baeck Institute was awarded one of the first grants in 
the newly established Metro Digitization Grants Program. The 
experience gained in this project enabled us to continuously enlarge 
our digital rare book collection. Working closely with scholars in the 



Digitizing the Rare Book Collection of the Leo Baeck Institute 187 

field and using feedback from our users in order to identify a 
meaningful selection, so far we have successfully digitized 167 rare 
books (about 22,000 pages) from our rare books collection. Among 
those books are 33 illustrated and art books which were funded by 
another Metro Grant awarded in 2008. By now we have worked out 
the technical details and workflows for our digitization program and 
are identifying new funding sources in order to continue the process. 
Newly digitized books are continually being added to our digital 
management system and made available on the Internet. 

Future projects: in 2008 the Archives at the Leo Baeck Institute 
embarked on an interesting new project - "DigiBaeck," which is 
destined to put all LBI archival holdings onto the World Wide Web in 
close cooperation with the Internet Archive. In this framework we 
consider to digitize parts of our book collection as well. Besides, we 
are collaborating with other libraries and have applied for shared 
digitization projects. 

Web interface and Examples 

The digitized books are accessible through the web interface of our 
digital management system Digitool (Ex Libris the Bridge to 
Knowledge, 2010). The books can be retrieved via a search interface 
or can be browsed by various categories. Searchable are the 
bibliographical data and the full-text, most of the books underwent 
OCR (optical character recognition). 

Additionally we reference the digital objects in our online library 
catalog records by adding the URL in MARC field 856 (see 856 - 
Electronic Location and Access, 2003.) 

Challenges 

This chapter focuses on the challenges faced and solved mainly during 
the implementation of LBI's digitization program in 2005/2006 as 
well as during follow-up projects. The problems were typical for 
projects which engage in new technologies, ranging from technical to 
organizational hurdles. Some of the difficulties which we encountered 
were connected to the special medium rare books. 



188 Digitizing the Rare Book Collection of the Leo Baeck Institute 

Challenge 1: Outsourcing a Digitizing Project 

Choosing the right vendor, developing a set of criteria for finding 
the right vendor including testing of sample files, negotiating a 
contract, monitoring and testing the results proved to be a very 
challenging endeavor. 

During the selection process for a digitization vendor, we 
developed the following set of criteria and asked vendors to submit a 
proposal with information and price quotes for the following areas: 

• Master files (TIFF) 300, 400, 600 ppi in bi-tonal, grayscale 
and color 

• Derivatives (jpeg, jpeg 2000) 300, 400, 600 ppi in bi-tonal, 
grayscale and color 

• Scanning oversized materials, foldouts 

• Technical and structural metadata 

• File naming 

• Cameras used 

• Book cradle 

• File storage and delivery 

• OCR (Optical character recognition) 

• Insurance 

• Security (handling of books, security of lab area, book tracking 
system, etc.) 

• Shipping and delivery 

• Quality control 

• Other costs 

• References 

We provided vendors with samples from our collection and asked 
them to provide us with scanned images and metadata. The scanned 
images underwent a quality check, the metadata was checked against 
standard recommendations. The best thing to do would have been to 
load the sample files into our digital management system, but at the 
time of our vendor negotiations our digital management system was 
not yet fully implemented. 



Digitizing the Rare Book Collection of the Leo Baeck Institute 189 

Overall we compared the vendor proposals and the samples very 
thoroughly, especially the price quotes for different scanning 
scenarios, the setups of the scanning labs, the quality of the scanned 
images, and the vendors' ability to produce structural metadata for 
our digital management system. The determining factors were the 
book cradle solutions and security features (e.g.: handling of books, 
security of lab area), the quality of the scanned images, the ability to 
provide complex metadata as well as the price differences. 

During the negotiation process we encountered variations of 
important technical details (e.g. file types, file naming, cameras used, 
technical and structural metadata, OCR) which helped us to make 
decisions which we then incorporated in the final contract. The 
contract spelled out our decisions in the listed areas. We learned that 
it is very advisable that the contract describes as many technical 
details as possible. 

Challenge 2: Medium Rare Bound Books from Previous 
Centuries 

One of the biggest challenge to digitize rare books from previous 
centuries is the fact that flatbed scanning is usually not an option 
since bindings cannot be removed for the scanning process. Rare 
books in our collection are usually treated as artifacts, the content is 
as important as the bindings and physical appearances. Older books 
are often very tightly bound and cannot be opened 180 degrees, it is 
very common that they can only be opened 90 to 120 degrees. 

Many of the scanning services which we approached were only 
equipped for flatbed scanning and declined to take on projects with 
bound books. Commercial solutions explicitly for rare book cradles do 
not seem to exist. Those few vendors we encountered which also scan 
bound books usually built or customized their own solutions for rare 
book cradles. Some used commercial book cradles for modern books 
which can be opened 180 degrees and added features to make them 
suitable for books which cannot be opened 180 degrees. Others 
vendors constructed adjustable boards to hold the books in place 
without applying pressure during the scanning process. 



190 Digitizing the Rare Book Collection of the Leo Baeck Institute 

The Grass Lipper Digital Laboratory at the Center for Jewish 
History designed and uses two book cradles, one cradle has a fixed 90 
degree angle, the other one a 120 degree angle. The cradles are made 
of acrylic and covered with black velvet. Velcro straps keep the book in 
place. The use of 2 "fixed" angle cradles turned out to be a simple, but 
successful solution. 

The different book cradles which we encountered represent 
different solutions for accommodating books which can only be 
opened at an angle less than 180 degrees. It is hard to make a general 
recommendation for a best solution. Depending on the type of books 
and condition of their bindings some cradles are better or less suited 
for the task of supporting fragile bindings and spines and of exerting 
as little pressure on the books as possible. For some books it is 
possible to use glass plates, but glass plates put too much pressure on 
most of the books. It is a trade-off - if glass plates cannot be used it 
very often results in shadows which often cannot be removed through 
the adjustment of the lightning. Overall those solutions which had a 
support for both front and back bindings were better suited for our 
specific collection. 

Rare books require additional manual handling time in securing 
the volume after each page is turned which makes the process more 
costly than scanning regular books. 

Challenge 3: Technical and Structural Metadata for Compound 
Objects: 

Not all scanning vendors are capable to provide the necessary 
technical and structural metadata for compound objects. During our 
search for a scanning service in 2005/2006 we realized that 
digitization vendors came from different backgrounds - 
photographers, microfilming companies, library services, 
bookbinders, etc. - and were thus often not aware of or capable of 
adjusting to emerging library best practices and standards. Many of 
them were able to deal with single digital objects and equipped to 
deliver the technical metadata, but not able to provide ready-to-go 
structural metadata in order to build complex digital objects. 



Digitizing the Rare Book Collection of the Leo Baeck Institute 191 

Our digital management system DigiTool (Ex Libris) requires the 
use of METS (Metadata Encoding & Transmission Standard, 2010) for 
storing structural metadata along with descriptive, administrative, 
technical and other kinds of metadata in an XML "wrapper." During 
our vendor search in 2005/2006 we found out that at that time many 
vendors were not aware of or capable to provide this type of metadata 
with the scanned images. If the scanning process is outsourced it is 
very advisable that the vendor provides structural metadata at the 
same time, otherwise it becomes a very tedious and time-consuming 
endeavor to match the digital objects to a structure map afterwards. 
During the testing phase we asked the vendors to supply us with 
metadata, preferably METS. During our initial project we could only 
find one vendor who could provide us with METS files. During the 
first phase of our project we additionally discovered that we needed a 
translation code between the METS ALTO version that the vendor 
could create and the METS ALTO version that our digital 
management system understood. 

Challenge 4: Organizational Parameters & Limitations of 
Digital Management Systems 

We had the "problem" that we had to use a digital management 
system which was chosen by our umbrella organization and which was 
not fully implemented at the time when we started our digitization 
program. The challenge was to bridge shortcomings of the system and 
shortcomings of what digitization vendors could provide, a typical 
problem when a technological field has not yet developed mature and 
commonly accepted standardization guidelines. The first versions of 
our digital management system did not accept structural METS files 
in METS ALTO standard, so we had to find additional funding for 
developing a translation code between METS ALTO (what Vendors 
were able to produce) and the METS "dialect" that our system 
understood, which resulted in higher costs. 

Organizational set-ups and the choice of the technical system 
have a great influence on the direction and the outcome of a project. 
These factors cannot be underestimated and they influence the 
timeline and the financial framework of a digitization project. 



192 Digitizing the Rare Book Collection of the Leo Baeck Institute 

Challenge 5: Keeping up with Evolving Digital standards 

One of the challenges we faced in follow-up projects was to keep 
up with evolving and maturing guidelines for digitization. One of the 
lessons learned was that it is necessary to spell out technical 
requirements in detail in a contract and to make sure that they are 
applied, and not to take them for granted. General formulations can 
lead to ambiguous interpretations. 

Example l: An area of dispute in one of our project was if it is 
"allowed" to enhance or sharpen master files during the scanning 
process. Many labs follow the NARA guidelines (see National 
Archives, 2004) and do not apply ANY sharpening or other 
"enhancements" during capture or to master files. If errors are 
detected during the capture, then the camera is calibrated. One of the 
vendors enhanced the master files instead of calibrating the camera. 
Our contract read: "Quality of image: Files will be inspected for 
sharpness, image contract, density and faithfulness to original and 
improved if necessary." - that should have only be allowed for 
derivatives, and not for master files. The master files should be true to 
the original materials; colors and contrast of the digital files should 
represent accurately the original pages. Many of the master images 
had too much contrast (visible in bad histograms) which probably 
came from sharpening during the scanning. In that respect our 
contract was not specific enough, we did not specify that only 
derivatives could be altered, but not master files. 

Example 2: It proved to be necessary in our 2008 project to 
spell out in the contract to check regularly if the calibration of the 
camera results in an accurate image. In our contract we had specified 
that the images produced should be "faithful to the original" 
materials. The colors and contrast of the digital files should represent 
the original pages accurately. 

Our first vendor calibrated the camera only once at the beginning 
of the scanning process to the "default" NARA guidelines. The results 
were aggravating: the majority of the image files had extensive color 
bleeding, a severe image quality problem. Ultimately we had to rescan 
our books at another digitization facility. 



Digitizing the Rare Book Collection of the Leo Baeck Institute 193 

Our contract was sufficient in the sense that we could rightfully 
claim that the bad image results were a breach of contract. But we had 
to rescan all our books - something which is definitely not advisable 
for rare books. A line in the contract about periodic camera calibration 
could have prevented this dilemma. We should have also done quality 
checks periodically. 

Lesson learned: The contract texts for projects in emerging 
technological fields are often vague and only give a narrative 
description of the wanted outcome since various schools of thought 
exist in terms of technical specifications. 

In many cases a narrative formulation in the contract about the 
desired outcome is the only possible way to address an area of 
concern, but in some cases it is not sufficient. The more hardcore 
technical details can be specified in a not ambiguous way the better. It 
is also very advisable to control the project frequently, to do several 
test phases, and not to rely on past vendor performances. 

Conclusion 

The successful implementation of our digitization program was and is 
based on a combination of persistence, serious initial testing, vendor 
comparison, ongoing testing of results, questioning of vendors, 
collecting of technical information, balancing of outside factors, 
changing of vendors when the outcome did not meet our expectations. 
The most challenging aspect of our rare book digitization project has 
been the steep learning curve of getting into the technical intricacies 
of creating digital content, and balancing outside factors to get the 
project on its way and moving. The most rewarding aspect has been to 
make hitherto relatively unaccessible books accessible to a broader 
audience. 

An important lesson learned was that is necessary to keep up with 
evolving and maturing guidelines for digitization - and to make sure 
that they are really applied, and not to take them for granted. Even 
established workflows have to be reevaluated from time to time, since 
the field changes all the time, and the technical details become more 



194 Digitizing the Rare Book Collection of the Leo Baeck Institute 

detailed and sophisticated. Digitizing of rare books is a very manual, 
labor, and ultimately cost intense work. 

References 

856 - Electronic Location and Access. (2003). Retrieved May 9, 2010 
from http://www.loc.gov/marc/bibliographic/bd856.html 

ExLibris the bridge to knowledge, Overview. (2010). Retrieved 
March 15, 2010, from 

http://www.exlibrisgroup.com/category/DigiToolOverview 

Metadata Encoding & Transmission Standard. (2010). Retrieved May 
9, 2010 from http://www.loc.gov/standards/mets/ 

National Archives. (2004). Technical guidelines for digitizing 

archival materials for electronic access: Creation of production 
master Files - raster images. Retrieved May 9, 2010 from 
http://www.archives.gov/preservation/technical/guidelines.html 



The Caprons of Paris: A Digitization 
Project in a Small Library System 



Misty De Meo (County of Brant Public Library) 



Abstract 

The County of Brant Public Library set out with the goal of 
documenting the history of the founder of the town of Paris by 
digitizing a collection of his original papers. By building contacts 
within the community, the Library was able to successfully complete 
its initiative and open new avenues for future projects. The equipment 
and digitization methods used are described, with a special emphasis 
on the methods in which the Library was able to circumvent its small 
budget. The Library's complementary local history wiki, and the 
project's methods of promotion, are also described. 

Keywords: Digital collection, Digitization, Historical society, Local 
history, Marketing, Web 2.0, Wiki 



Introduction 

One of the key goals of the County of Brant Public Library is to act as a 
gateway providing the most accessible routes to information. The 
County of Brant Public Library Digital Collections was envisioned as a 
means by which local historical information which was currently 
unknown or inaccessible to most members of the community could be 
made easily and publicly available. The goal of this project was not 
merely to provide information, but to provide the means by which the 
community could interact with history and share information. 



195 



196 The Caprons of Paris 

The County of Brant, located in southern Ontario, is a diverse 
collection of unique communities, each with its own history. It has a 
modest population of 34,415 divided among 30 distinct communities, 
formerly separate municipal entities, spread out over 843 km 2 
(Statistics Canada, 2006). The Library has five branches spread 
throughout the County. It was impossible for a single project to 
document the County as a whole. This digitization project was 
designed as the first part of an ongoing digitization program which 
would eventually represent the entire "community of communities" 
that makes up the County. The short-term goal was to document the 
history of the town of Paris, the largest community in the County of 
Brant, through its founder Hiram Capron; the long-term goal was to 
build an ongoing program documenting all of the communities in the 
County, using the Capron project as a method to open doors and 
create momentum. 

The keys to the success of this project have been technical 
flexibility and innovation, which enabled success within a modest 
budget, and community engagement, which provided the support, 
contacts, and materials necessary to grow a small project into a pair of 
ongoing initiatives. 

Project selection 

The Library evaluated a number of potential project partners when 
planning this project, from communities across the County. 
Unfortunately, for various reasons, we were not able to work with all 
of the local historical societies we approached. 

Most importantly, we needed a critical mass of original materials 
- a body of materials large enough to form a hypothesis as the basis of 
a project. Local history is exponential - pursuing a sufficiently large 
collection of materials leads to new collections and new potential 
partners. While many potential projects had seemed promising, a 
number proved to be unsuited to this project on examination; some 
collections were small and lacked that force behind them which we felt 
was necessary to give us an "in" to the community, while others 
proved to contain many inauthentic items and reproductions which 



The Caprons of Paris 197 



were unsuited to this project, which was focused on original primary 
historical documents. 

An ideal project was eventually found in the Paris Museum and 
Historical Society, whose archives hold a large body of original 
documents. Hiram Capron, the town's founder, is an important figure 
in the history of Paris and is almost unusual in that his legacy is by 
and large authentically supported by the documents the museum 
holds. The museum was willing to consider a partnership, and 
provided the Library with open access to its Capron collection. At the 
time of the agreement, approximately 50 items were identified for 
digitization in the project. The "critical mass" criteria proved apt: at 
the time of writing (March 2010) the Library has digitized a collection 
of over 425 items, provided by the museum and by new contributors, 
as a part of the Hiram Capron project and further projects. By 
beginning with the founder of Paris, we have been able to broaden the 
perspective to provide a wide variety of materials about the town's 
history and to forge partnerships with new donors. 

Several historical societies have expressed concern of retaining 
ownership and control over their items. One of the necessary keys to 
developing partnerships was to show potential project partners that 
digitization could enhance the value of their collections and their 
services, rather than replace them. Our partners at the Paris Museum 
and Historical Society have reported that their visits have increased 
since the launch of the digital collections websites. 

Management and staffing 

The project involves three key members of the Library. The Library's 
CEO, Gay Kozak Selby, originated the project and secures funding 
from outside sources; she performed initial project planning, research, 
and hiring with the library's e-resource librarian, Christine 
MacArthur. Christine was also responsible for day-to-day 
management and co-ordination of promotion. Scanning, technical 
management, and design of promotional materials was the job of the 
professional archivist, Misty De Meo. 



198 The Caprons of Paris 

In addition, the project has made use of shorter-term student 
workers in order to handle particular extra tasks and features. In the 
summer of 2009, the Library employed a university student using 
funding from Young Canada Works to provide transcriptions of a large 
number of handwritten documents in the Capron collection in order 
to make the full text searchable online. Other projects of this nature 
are anticipated in the future. 

Funding 

Funding was provided through the Library Strategic Development 
Fund, an ongoing grant operated by the Ontario Ministry of Tourism 
and Culture with the goal of assisting innovative projects and 
community development. The Library was one of the recipients for the 
2008-2009 grant, receiving $18,405. This grant was critical to the 
success of the project; it could not have gone forward without external 
funding. It provided for equipment costs, the hiring of a professional 
archivist to supervise and perform digitization, and some of the 
project's operating costs. The total operating budget was $35,000, 
with the remainder of the budget paid by the Library. Summer student 
work has been funded though Young Canada Works, a program 
operated by Canadian Heritage. The Library's primary website was 
provided by a grant from Knowledge Ontario. 

Content Management System 

A number of factors influenced the selection of the content 
management system used by the host for the digital collection. These 
included ease of use and intuitiveness of user interface; advanced user 
interface features; searchability; and accessibility through external 
search services, such as Google. While the Library holds a license to 
use SirsiDynix Hyperion, we felt that it was insufficient for our needs 
after evaluating a neighboring institution which had used it (King 
Township, 2008). Its interface is somewhat cumbersome and 
simplistic; we decided that its "look" would turn off a large number of 
visitors. It also lacks advanced features for object descriptions or 
supplementary viewing options, such as contextual maps or 



The Caprons of Paris 199 



"zoomable" image views. Perhaps most importantly, its results are not 
available through Google - we recognize that most of our potential 
visitors will be using tools other than our own built-in search engine 
to discover content, and so leaving our content out of Google 
needlessly keeps our content away from interested visitors. 

In the end we selected a website toolkit called VITA created by the 
Our Ontario division of Knowledge Ontario, a local organization which 
provides a variety of digital services and content. Our Ontario, which 
focuses on providing access to digitized historical materials, provides 
grants to small organizations and so we were able to obtain use of the 
software and hosting free of charge. 

Our visitors have generally been impressed with the advanced 
features VITA makes available. One of the most popular features is its 
Google Maps integration. Each item can be tagged with a set of 
coordinates that will cause a Google Maps widget to display below the 
item. We have used this extensively with items such as historical 
maps, which allows us to display a modern image of the town centered 
on the same location as the map. Another popular feature is the 
integrated "Zoomify" software, which allows the user to zoom in to an 
oversized item at greater levels of detail, or to pan and rotate. An item 
which demonstrates both of these features is available at 
http://images.ourontario.ca/brant/details.asp?ID=68322 

VITA is also designed around modern Web 2.0 search methods; 
consequently, in addition to its own built-in search, it makes all of its 
content available via Google and permits material access and sharing 
through RSS and social networking sites (Knowledge Ontario, 2009). 
Thanks to this, our VITA site features prominently on the first page of 
Google results for key terms such as "Hiram Capron". Our site makes 
significant use of user-interaction features, encouraging users to 
comment and contribute information. Many items contain "mystery 
questions" inviting users to submit information that is missing about 
an item. 



200 The Caprons of Paris 



Our Brant: local history, local voices 

The project, as originally planned, comprised only one digital history 
site displaying primary documents. Since then, we have made the 
decision to launch a second local history website at the same time as 
our primary site, currently available at http://ourbrant.wikia.com/ As 
materials and contributions for our primary site were being prepared, 
it became clear that a great deal of information would not be 
appropriate for inclusion in the primary site. As is almost certainly the 
case with most established communities, much of the history of Paris 
has by this time become a matter of tradition which is not entirely 
substantiated by the remaining documentation. 

The goal of the primary Digital Collections site was to provide 
original historical documents, photographs and other items in a 
digital format, both for the purposes of historical research and to 
allow people to learn about the town's history. For this second 
audience, it was especially important to provide peripheral materials 
which slotted each item into the "narrative" of historical events in 
order to build understanding, which was achieved both through item 
descriptions and through explanatory photo essays. However, because 
the site is focused on the documents and historical evidence more 
than the story, it has been considered very important that the 
narrative presented is not at odds with the historical documents and 
that it does not make assertions which are not supported by the 
documents. We were further interested in enhancing interaction with 
our users and clients; we felt that the Web 2.0 style interactive 
features in the VITA website software was one of its greatest 
strengths, and we looked for further opportunities to enable this kind 
of interaction within our communities. 

As a result, it was decided to add a second site to the project. The 
goal was to provide an appropriate location for people from the 
County to document the area's history as they personally understood 
it. With a focus less on hard historical facts and records, it would 
provide an appropriate place to record these popular anecdotes of 
local history. Hence, we defined the primary goal of this site as 



The Caprons of Paris 201 



providing a place for anyone from the county to share their personal 
histories. 

Based on this goal, we decided to provide spaces for the following 
types of content: (1) Personal memories and life stories; (2) Family 
histories; and (3) Profiles of notable local residents, buildings, and 
organizations. 

The wiki format was judged to be most appropriate for this 
application due to the ease with which it facilitates collaboration and 
open contributions with a relatively low barrier of entry for new users. 
The wiki format was also judged appropriate due to the abundance of 
services which provide free hosting, such as Wikia 
(http://www.wikia.com/), as this would eliminate a financial barrier 
to the creation of the site. 

While we have found other local history wikis during the planning 
process, we have not found another site with the same focus and 
consequently much of Our Brant's design is entirely original. Many 
sites we have found focus on locations as grounding points for 
memory, such as Placeography (http://www.placeography .org/) and 
Zurbu (http://zurbu.net/), while a few others, such as the Wagga 
Wagga Local History wiki (http://waggalocalhistory. wetpaint.com/), 
focus on events. Furthermore, a number of these wikis are not open to 
the public for editing, such as the Alexandrina Local History Wiki 
(http://alexhistory.pbworks.com/) or the Montana History Wiki 
(http://montanahistorywiki.pbworks.com/). Our Brant differentiates 
itself from these other services by providing open access for editing to 
any registered or anonymous users, and by placing a strong focus on 
individual history and personal memory rather than using the anchors 
of locations and subjects. 

Input Methods 

Ease of editing was considered to be one of the most important 
features. Many of the residents in Paris and the County who are 
interested in sharing local history are not very comfortable with 
computers. Despite the familiarity of Wikipedia, it is not simple to 
create pages without using a special markup language (Baker, Hoover 
& Rose, 2009); given our probable audience, this was not considered 



202 The Caprons of Paris 



acceptable. While Wikia includes a rich text editor which abstracts 
markup language from the user on individual pages ("Rich Text," 
2009), it does not automate elements such as page structure or link 
structure. We chose to adopt the forms-based input system used by 
Baker et al. on Placeography, which presents the user with a simple set 
of input boxes and checkmarks using software called Semantic Forms. 
Once submitted, the result is an attractive, professional-looking page 
without requiring the user to employ any markup language. 

Another benefit of the Semantic Forms software, and the 
Semantic MediaWiki software on which it is based, is the ability to 
perform queries on the information entered in form fields (Semantic 
MediaWiki, 2009). Our Brant uses this extensively to build the index 
pages which allow users to browse the site's content. Because these 
indices are based on queries, they update in real time to include newly 
added content; this has simplified page creation by eliminating the 
need for users to create links to their pages (e.g., see Figures PARIS -1 
to PARIS-3). 

The community pages provide the central browsing interface to 
the site's content, using queries to aggregate together all content 
which is marked as belonging to this community. Each community 
page also contains links to the forms for creating new pages, and each 
individual page contains an "edit with form" link which allows it to be 
altered using the same form interface. 

Participation 

Baker, Hoover and Sherman (2009) note the importance of 
building a community in order to create a successful wiki, and we took 
their lessons to heart when preparing Our Brant. We approached local 
community members about contributing to the site prior to its launch; 
these included the president of the historical society we partnered 
with as well as stakeholders in the area's history, such as the owner of 
the historic Asa Wolverton House in Paris. In addition, we drew on 
our own resources and digitized transcripts of oral history interviews 
that the Library had conducted in the 1970s and 1980s. This allowed 
us to present the site at launch with a significant amount of content to 
attract users and to provide examples of what they could contribute. 



The Caprons of Paris 203 



ujikia 


Create a new wiki 




Edit with form f Edit this page ^History ©Delete BlMove 11 Protect v' = Unwatch C-Refresh 


Edit Place: Asa Wolverton House 


Address: |52 Grand River St, S. 


COUNTY OF BRANT 

^uhlic T ilirarv 

L UU11\. LiiUlUl J 


Located in: | Paris _jj 
Built in: |l851 
Renovated in: 


| Search Our Brant 
Our Brant 

Communities > 
Explore the wiki > 


Original function: | House _jj 
Current (unction: Bed and breakfast 

Photos: To upload s photo related to (his place, click "add another" below. Click here for help with adding photos. 


Contribute ) 
Memories ) 


Photo: |Asa Wolverton portrait.jp cj Upload image 


Families > 


Caption: [Portrait of AsaWolverton with miniature 


Places > 
Mew pictures 


Remove | 


Site Community > 
^/ Create a new article 
Jj Upload .1 new image 

Recent changes Random page 
Help Special pages 


Photo: jAsa_WolvertonJiouse_door.jpg Upload image 
Caption: Plantation shutlerfront door 
Remove | 




Latest activity 1 


Photo: lAsa_Wolverton_house_sketch.jpg Upload image 


8*1 articles on this wiki 


Caption: Larger copy of sketch 


& Asa Wolverton House 

3 seconds ago by ■ %derneo 

County of Brant 

20 hours ago by Mistyderneo 

if OurBraii1:Tu1o rials 

3 days ago by Mistyderneo 


Remove | 


Add another | 


Complete description of this place: xss. Wolverton emigrated to Paris in the 1830 3 from Cayugi 
a sawmil 1 soon after: his arr ival . This enab led him to de; 

Tim]^p- r.lip 1-T.rn np i nVihnur i nn hniifip^ t.n r.lip Pinur.h nnrl nnir.p 





Figure PARIS-1 Sample form for creating a page for a building 



Asa Wolverton House 




T tiers are fivf original firs places in trie hoo: e low lain ■■ a isceiitlj j king filer rhe kitcliei here wae connecter! 10 the dining room Uni*ntfi.in<:.ti-M) Eeo ana isreakfasi 




Memories about Asa Wolverton House 




Figure PARIS-2 The result page 



204 The Caprons of Paris 



Memories 


People 


■ John Anderson 


■ Syl Apps (hockey player and 


■ Mrs. Ernie Buck, Marge McCorkindale, Mr. Be Mrs. 


Conservative MP) 


Andy Leishman, Alice & Andy Scott 


■ Hiram Capron (founder of Paris) 


■ Edith Carnie 


■ Walter Capron (Hiram Capron's nephew) 


■ Tom Cocker 


■ Alma Duncan (artist) 


■ Clara Farr 


■ Ron Eddy (current mayor of the county) 


■ Gord Gibson, councillor 


■ Hugh Finlayson (first mayor of Paris and 


■ Eleanor Hardie 


tanner) 


■ Bob Hasler, president of the Paris Museum and 


■ Norman Hamilton (prominent early 


Historical Society 


businessman) 


■ Margaret Laing, librarian 


■ John Penman (owner of Penman's Mills) 


■ Norma Leighfield, reeve 


■ Ted Reader (television chef) 


■ Muriel and Jack Pickell, owners of the Paris Star 


■ Jay Wells (hockey player) 


■ Corydon C. Randall 


■ Robert West (boat renter, entertainer, 


■ Jackie Remus 


and eccentric) 


■ Kay Riddolls, nurse 


■ William West (poet and owner of the 


• DaleRobb 


Nith Navigation Co. boat rental 


■ Foster Scott, teacher 


company) 


■ Mel Sharpe, mayor 


■ HriaTloc TAThit"]aw f Trillion ann tti a ^rf^T^ 

■ ^11^1 ^^IIILIQVV gunnel aii^-i Lll&.y*jL J 


■ D.A. and Isobel Smith 


■ Paul Wickson (artist) 


■ James Thomson, Penmans employee, and wife 


Click here to add a person if you don't see 


■ Police Chief Watts 


someone on the list. 


■ Shirley Williams, from Qua family 




■ Leonard Wise, owner of Harry Wise Menswear 


Places 


Click here to add your own memory about this 




community! 


Churches 




■ Paris Baptist Church 


Family histories 


■ Paris Presbyterian Church 


■ Patton family, owners of Elton Cottage 


■ Sacred Heart Roman Catholic Church 


^ t j. _i: rii -l. 



Figure PARIS-3 The browse index for Paris. (Shortened, as the complete list is too 
long to be shown.) The lists of memories and of people are generated by queries and 
are updated automatically whenever a new page is created. 



This proved to be a success; when examining our statistics, we 
found that visitors had discovered our site even before the launch and 
were sharing links to our content online (Qua, 2009). In addition to 
complete content, we have also seeded the site with a large number of 
"stub" pages - short articles with only a minimum of content, which 
contain invitations to the reader to complete them. 



The Caprons of Paris 205 



In order to help promote usage of the site to an audience that may 
not have access to the internet, we have also started a program to 
accept site contributions on paper; this ensures that those who do not 
own computers or who do not feel comfortable with computers can 
still contribute to the site. 

Digitization standards 

Our digitization standards were based on those recommended by 
a number of sources, primarily the U.S. National Archives and 
Records Administration (2004), and the Canadian Museum of 
Civilization (2007). Due to the size of our organization we were not 
able to completely adhere to these conditions; very specific standards 
pertaining to digitization room design and monitor calibration are not 
realistically achievable using the budgets of most small organizations. 
Our standards include both a web display format and a preservation 
format, which is meant to remain accessible and usable in the long 
term. 





Preservation masters 


Display copies 


Format 


8-bit TIFF (from scanner) 
DNG camera raw (from 
camera) 


8-bit JPEG 


Resolution 


600dpi (from scanner) 


Minimum 1000 pixels along 
longest end; higher when 
necessary 


Colour space 


sRGB IEC61 966-2.1 


sRGB IEC61 966-2.1 



We selected the Digital Negative (DNG) format as a master 
format for camera images instead of TIFF because of the increased 
flexibility it offers (Adobe, 2009). While proprietary "raw" formats are 
archivally unsound, because of their closed specifications, DNG offers 
a manufacturer-neutral format for this data. Adobe has made the 
specification publicly available, which ensures that it can be supported 
by future software. There is no archival consensus on the use of DNG 
as a preservation format (Hess, 2009), but we feel that its benefits 
outweigh any risks. When storing preservation copies on media 



206 The Caprons of Paris 

without significant limitations on storage space, such as the County's 
central hard drives, we also store uncompressed TIFF copies of 
camera images. 

Our standard workflow is as follows: 

• Import image from scanner or camera raw into Photoshop as 
16-bit RGB, using the sRGB profile 

• Perform any necessary colour correction and cropping 

• Convert image to 8-bit colour depth 

• Enlarge image by 0.5 inches or 1 inch 

• Type item number in lower right corner, and partner 
information (if appropriate) in lower left corner 

• Save preservation TIFF 

• Resize item to display size and perform sharpening 

• Save web JPEG 
Books 

A special note should be made of our standards in digitizing 
books. Many digitization projects, such as Google Books, use 
computer processing to flatten pages, remove bindings, and recolour 
pages to a bitonal, or pure black and white, format. We chose to use a 
different presentation method. This digitization project is primarily 
archival in nature, and most of the bound material selected was 
chosen for its archival qualities as much as its informational qualities. 
A guiding rule of archives is that context is as important as content in 
reading meaning from an item (Van Ballegooie & Duff, 2001), so our 
standards are designed in order to ensure that the context of a book's 
usage is evident in our digitized version. 

A good example of this is our digitized copy of Frederic A. 
Holden's 1859 Genealogy ofBanfield Capron, a family history of the 
Caprons. The copy available to us was Hiram Capron's personal copy, 
which was later owned by a succession of other Caprons who have 
made extensive annotations and additions to the book. 



The Caprons of Paris 207 



C A P K 0 N FAMILY. 



107 



[901] 



Sarah M. Capron, daughter of Otis and Polly, was bom 
August 22, 1796, and married Samuel Boyden, July 19, 1824. 

CHILDREN. — VI. GENERATION. 

977 Samuel Boyden, Jr. ; born May 28, 1825 ; died young. 

978 Sarah Curtis; born August 7, 1827. 

979 Martha Maria ; born November 11, 1829. 

980 Samuel, Jr. ; born April 28, 1881. 

981 Benjamin Franklin ; born September 4, 1833. 

982 Mary Elizabeth ; born May 24, 1835. 

983 Edmund Capron ; born September 1, 1838. 



Samuel Boyden, Jr., sou of Samuel and Sarah, married 
Ellen L. Morse, October 17, 1855. 



Benjamin Franklin Boyden, son of Samuel and Sarah, 
married Maria Louise Kingsbury, November 28, 1857. 



Judith Capbon, daughter of Otis and Polly, was born Dee. 
13, 1801. She married Nicholas Cook. 



[ 980 ] 



[981] 



[ 903 ] 



CHILDREN. — VI. GENERATION. 




Digitized by 



Google 



Figure PARIS- 4 Page 107 from Google Books's edition 



208 The Caprons of Paris 




Cap i«i miLr. 



107 



[9*1] 



tin 



^S***lt M. OUMK, J- .^jbler of CHu and l'.JIr, ni born 
I . Sliu^rt U12, LT'."i. al,J Hgnjixl grand BojJ^ Jidj 19, lflii 

J it »V ■ 




I m»rrH=3 



TSmjkis I'mmir Bemss. eon uf Samnsl ami 8u»li. 
■umtx] Marii LmiiM Kingiib.rj, N<,™mh«, 2^ 1957. 




jLijltu C*»usw. iiHRhlf-r <,f Oci* ind *u km Dm. 

IS. 1801, t*« dhttM NieheJu C«oV. 

CJIJL.D1LKX- VI. (iESERATTOS. 

LmiiiA <'i»ik : msrriivl Alrin 
HO ; nairriwl Jiraot lunik 

Minia ; married Mary alartift. 
«»T llanuh ; nitwl '<*■ Knwlnll. 
»«* AI.Mlli . : uiarr.nl Lovull I'm k.-r! :i ■. 
»a Xichata* Jr, * 
JuJilh. 



! 



Property of Paris Museum 



Figure PARIS-5: Our edition of this page, showing the original book's condition. The 
page behind the attached card was also digitized separately 



The Caprons of Paris 209 



These annotations are as important for our purposes as the 
original book. They identify some of its owners, and provide 
contextual information on how this family history was used and why it 
was important to them. For that reason, we have chosen to digitize 
these archival books in full color with any annotations or additions, 
maintaining the full size of the page and bindings. 

Metadata 

To simplify metadata collection given our resources, the Library 
has focused on using the VITA software's metadata for our digitized 
items. Metadata is entered at the time a record is prepared for display 
in VITA, and is preserved by saving local copies in the VITA and 
Dublin core formats. VITA's metadata is stored in an XML-based 
format, which allows translation into other formats as needed in the 
future. Metadata is standardized using a number of mandatory and 
preferred fields, and supplemented with any additional relevant fields 
per item. 



Mandatory: 


Preferred: 


Contributing Partner 


Citation 


Copyright Status 


Collection 


Language of Description 


Copyright Date 


Media Type 


Copyright Holder Name 


Title 


Copyright Statement 




Creator Name 




Date(s) of Original Donor 




Earliest Year 




Latest Year 



Preservation 



Both the preservation master copies and the web copies are 
preserved, along with metadata and any OCR/transcription data. Two 
physical copies are created on archival quality DVDs; one set is kept 



210 The Caprons of Paris 



by the Library, and the other by our project partner. In addition, 
supplemental preservation copies are stored on the County's servers. 

Equipment 

The original project plan called primarily for the digitization of flat 
documents, which consisted primarily of single pages or a small 
number of pages, and the project's equipment and software was 
purchased with these types of documents in mind. The original 
equipment purchased for this project consisted of one computer 
workstation, with Adobe Photoshop CS4 software; two scanners; and 
general archival supplies. 

Scanners 

The Epson GT-20000 (Epson), the first scanner purchased, is 
advertised primarily as a workgroup scanner, and its primary purpose 
is scanning large size documents. It supports documents of a size up 
to 11.7" x 17", at a resolution of 600dpi. It was purchased on the 
recommendation of our partners at Our Ontario. 

The V500 (Epson) is intended primarily as a photo scanner; it 
was purchased as a secondary unit. It supports resolutions as high as 
6400dpi, with documents up to 8.5" x 11.7" in size, and unlike the GT- 
20000 it is able to scan photo negatives - this was the primary reason 
it was chosen. Both scanners support colour depths of up to 48-bit 
colour. 

Book scanning 

As the project progressed, new material came forward and it 
became clear that bound books and book-shaped materials would also 
need to be scanned. The majority of these items were Hiram Capron's 
personal account ledgers dating back to 1828, and were very fragile. It 
was clear that their brittle spines would not survive the bending 
necessary to flatten the book for use in a flatbed scanner. In addition, 
the natural curvature of ledger and book pages when laid out flat on a 
scanner would produce sub-optimal images, especially when 
performing OCR (Clements, 2009). Another concern was the scanning 
of oversized objects, as many items in the collection included elements 



The Caprons of Paris 211 



which were larger than the 11.7" x 17" size that the largest scanner 
could image. 

We decided to use photographic imaging for these items, which 
would free us from physically fitting items into a flatbed scanner. 
While we were familiar with commercial book scanning cradles, such 
as those produced by Atiz (Atiz) and Kirtas (Ristech), our equipment 
budget for this project did not permit a purchase of this expense. 
Consequently, we constructed our own simple cradle at a low cost. 

The only supplies required for the cradle itself were: 

• Two white foldout foam presentation display boards: 
Presentation boards proved to be ideal because they are 
segmented and naturally fold out. This allowed the longest ends 
to be raised at a 90-110 0 angle from the surface of the table, 
providing the surface for a V-shaped cradle. 

• Weights to hold the boards in place: The boards need to be held 
in place using physical weights of some sort, both to provide 
support to hold the body up at an angle and to prevent slippage 
from the weight of the book or artifact being digitized. Weights 
were easily produced from salvaged supplies. Binders were 
initially used, but these proved insufficiently heavy to guard 
against heavy objects slipping. Boxes of books from storage were 
ultimately the ideal solution, and certainly plentiful in any 
library. 

One camera, with an SD card: A discussion of the selection of the 
camera is contained in the next section. 

One tripod 

One plate of glass to hold down pages (optional): The use of 
glass permitted curling pages to be held flat while being 
photographed. However, because a polarizing filter was not 
available, glare prevention is a challenge. The plate of glass was 
taken from an unneeded spare picture frame. 

• Two large sheets of black paper, providing a scanning surface for 
items to rest on 



212 The Caprons of Paris 




Figure PARIS-6: The book cradle. 

The total cost of the supplies needed came to less than $700, and 
many of the supplies were already available at the Library. The only 
parts which needed to be newly purchased were the camera and its 
memory card, at a cost of approximately CAD $550. 

The photograph above depicts how the setup functions. The book 
or artifact being digitized is placed in the centre of the cradle. Next to 
it, the camera is placed on the tripod and angled facing down at the 
page. Because only one camera was used, only the even or odd pages 
are digitized at a time; after completing one pass, it is necessary to flip 
the book and photograph the other pages. After photographing, the 



The Caprons of Paris 213 



images can be transferred to the computer for processing. The use of a 
consistent angle and book positioning means that the images can be 
automatically processed using Photoshop batch processes, so it is not 
necessary to manually crop individual pages. 

The cradle provided significant advantages when compared to 
other options. It allowed books and ledgers to be digitized clearly and 
legibly while providing pages which are flatter in appearance than 
those scanned using a flatbed scanner. The use of a camera also meant 
that capturing each individual image was significantly more efficient 
than using a flatbed scanner, because camera imaging is much faster 
than flatbed scanners. In addition, the use of presentation boards 
meant that the size of documents could be very large - as large as 36" 
x 24". The greater limit on item size was camera resolution. The use of 
such a large surface meant that the cradle was also useful for digitizing 
other oversize archival records which could not fit in the flatbed 
scanners. 

When selecting a camera for use with the cradle, a number of 
criteria were considered. The criteria which guided our decision were: 

Resolution. While most high-end digital cameras have more 
than sufficient resolution to digitize average-sized books and items, 
the collections being digitized included some very large objects 
containing fine-grained detail such as a 19th century atlas that 
measured 15" x 17.5". These items necessitated a camera with a very 
high megapixel count in order to produce legible images. 

Raw compatibility. The "raw format" is a type of "digital 
negative" (Toborg, 2009), which allows cameras to record information 
exactly as received by the sensors; this enables more advanced post- 
processing (Canon, 2008). The most valuable feature this permits is 
quick and accurate colour correction. 

Cost. Any camera selected had to fit within the equipment budget 
remaining for the project, which eliminated DSLR cameras of a high 
resolution. 

Compatibility with professional book digitization 
equipment. If a future budget permitted the purchase of 
professional book digitization equipment, it would be most beneficial 



214 The Caprons of Paris 

to be able to use the camera which had already been purchased in 
order to reduce costs. 

DSLR (Digital Single-Lens Reflex) cameras are popular choices 
for book digitization (Torborg, 2009), particularly because of their 
high imaging quality and higher dynamic range (Wan, 2008). 
However, for this project another camera was selected which better 
met our needs. In particular, the requirement of a high megapixel 
count ruled out most entry-level DSLR cameras. Based on these 
criteria, we selected the Canon PowerShot G10, at the time the 
highest-end "prosumer" Canon PowerShot camera available. With its 
14.7 megapixel sensor 1 (Canon, 2008), it was capable of capturing 
sufficiently fine detail in large items such as the atlas. It also met the 
other criteria because it supported the raw format. There were no 
DSLR cameras within the budget available with a comparable 
resolution; the resolution was considered an acceptable trade-off for 
image quality and benchmarks indicate that, for the purpose of this 
project, the G10 is within an acceptable quality range compared to 
similar DSLRs (DxO Labs). It provides excellent image quality for 
digitizing books and documents even in non-ideal lighting conditions. 

Lessons learned from the cradle 

While the cradle has been very successful for imaging documents 
unsuited for flatbed scanners, determining the best shooting methods 
has been an ongoing process and both materials and practices have 
been amended as the project progresses. We have switched to a 
carefully configured camera setup instead of its uncalibrated default 
settings, and have improved the cradle's design. 



1 The G10 has recently been replaced by the Canon PowerShot Gil, which has a 
significantly reduced resolution of only 10 megapixels - a 33% reduction (Butler, 
2009, p. 1). While the reduction in resolution was for the purpose of reducing the 
amount of noise in images and increasing detail at lower resolutions (p. 17), the G10 
performs better and produces more detail at the lowest film speed (pp. 11, 14, 19) 
used in digitization. While the G10 was a suitable replacement for a DSLR for this 
project, the Gl 1 would not be. 



The Caprons of Paris 215 



There have been some minor tweaks to its physical construction, 
which have helped to significantly improve quality. The presentation 
boards work very well as an inexpensive, readily available material. 
White presentation boards were initially used because they were the 
most readily available medium. However, the Canadian Museum of 
Civilization (2007) recommends using black or neutral grey 
backgrounds for photography. In our testing, we determined that 
using a black background does produce superior results, with lower 
noise levels and a superior contrast ratio. 

More detailed accounts of our findings, and those of others, can 
be found at the DIY Book Scanner website at 
http://www.diybookscanner.org/ and at the archivist's personal blog, 
located at http://www.mistydemeo.com/ . 

Software processing 

As Wan (2008) notes, the Gio's image tends to be noisy direct out 
of the camera, and this is especially noticeable when the camera is 
shooting in its JPEG mode. In our experience, noise can be minimized 
and detail maximized by shooting using the raw format in well-lit 
environments; using appropriate exposure settings; using the lowest 
ISO ("film speed") level available; and by using Adobe's Camera Raw 
software for noise removal (included with Photoshop) in place of the 
Canon Digital Photo Professional software bundled with the camera. 
As with the white balance settings, software noise correction 
processing can be automated and hence does not adversely add time 
to the processing workflow. 

Promotion 

Both of our sites have been promoted using a variety of methods, 
including traditional methods such as print and through special 
events, and through non-traditional online methods. 

Online 

Our passive advertising methods have been primarily focused 
online and have centered around findability - ensuring that both of 
our sites can be located using search engines such as Google, and in 



216 The Caprons of Paris 

other places where users may be looking for related information. Our 
sites are linked and indexed in a variety of relevant locations, 
including our primary library site and the sites of local historical 
societies. These have helped to increase its Google ranking, ensuring 
that key terms such as "Hiram Capron" and the names of individuals 
for whom we have memories appear among the first search results. As 
well, we have included links on highly trafficked online resources 
which do not contribute to our Google ranking, such as the Wikipedia 
articles for Hiram Capron, Paris, County of Brant, and others. Our 
statistics show that, in the month of November, approximately 10% of 
our visits came through Wikipedia and 20% came through search 
engines. 

Interlinking between our two sites has also helped to direct 
visitors to our other materials; 25% of our visits on one site in the 
month of November came through links from the other. 

Press 

Press releases were distributed to the local media at various 
stages of the project, beginning shortly after the digitization work 
began; a new press release was distributed every month or two 
months. They provided ways of announcing the project and providing 
the community with short updates on progress, and also provided 
ways to communicate project events to community members. A week 
prior to the project's official launch, we also purchased a set of 
advertisements in all of the area newspapers. 

Local community newspapers were the most amenable to working 
with us. All of our press releases were printed by the local County 
newspapers, and the community newspaper from Paris sent out a 
photographer to document our Digitization Days event. 
Unfortunately, obtaining coverage and advertising in the larger 
newspaper from a bordering city proved more difficult; not all of our 
press releases were printed, and advertising space was substantially 
more expensive. 

We have found that the rumors of the death of the newspaper are 
greatly exaggerated. Print was one of our most effective advertising 
methods. Through informal questioning, we determined that most of 



The Caprons of Paris 217 



those contacting us about the project had read about it through 
newspapers. Our advertising space has also proven to be effective. Our 
Google Analytics report on the site's first week available to the public 
showed a substantial spike of traffic, representing our highest daily 
visits to date, immediately after our newspaper advertisements were 
printed. 

Close to the time of our launch and after it, growing interest in the 
project has enabled us to take part in local television and radio; this 
has helped promote the project and generated further interest. 

Advertising handouts 

In addition to our print materials, we prepared a number of 
advertising materials to distribute to our partners and directly to 
community members. 

To advertise our website and specific events, we designed three 
eye-catching posters. Two are permanent posters, designed to 
generate ongoing interesting the project, while one is a one-time 
poster intended to advertise a specific event. 

The permanent posters were created to advertise both of our 
websites. The primary poster was intended as a project-neutral 
advertisement for the site; consequently, it was designed to avoid 
specific references to the Capron or Paris projects and image selection 
was necessarily constrained to images which were not distinctly 
Parisian. The second poster is intended to solicit contributions to the 
Our Brant site, and specifically targets an audience which is less 
familiar with computers. In both cases, we were able to draw from 
photographs in the collection in order to obtain attractive design 
elements which could generate further interest in the collection itself. 

Due to the expense of having posters professionally printed, we 
opted not to design complex posters for most one-time events such as 
the Digitization Days event. However, we judged our launch to be 
important enough to warrant a professionally printed poster. In 
addition to relatively wide distribution to generate interest in the 
launch event, it was especially valuable for the purpose of distribution 
to the agencies which provided our grant. 



218 The Caprons of Paris 

The three posters were displayed in all of the branches of the 
Library. In addition, they were also distributed widely to project 
partners and other locations such as local genealogical societies, 
museums, and businesses. 

We also produced a set of promotional postcards to be distributed 
directly to Library patrons. Both the front and backs of the cards 
contain information on the websites with their URLs. The cards are 
prominently displayed at the circulation desks of all Library branches 
as well as our project partner, and clients are invited to take as many 
as they would like. 

Events 

The Library has held two special events in order to help build 
interest in the project. The first was a community event called 
Digitization Days, which was held on November 13th and 14th, 2009. 
It was conceived for the dual purposes of sounding out any additional 
material for our project which might be held by community members, 
and project promotion by digitizing individual families' photographs 
and documents to build interest in digitization and our website. 
During this event, community members could bring in their personal 
photographs or documents to be scanned and, for a small fee, could 
take home the scans on an archival CD. 

Much to our surprise, we found that the majority of the 
contributors were not interested in taking home any personal digital 
copies of the photographs they had brought to be digitized. They had 
brought their material specifically for the purpose allowing it to be 
included on our website. In total there were eight contributors over 
the two days of the event; of these, five brought material specifically 
for our website, and one other was willing to allow his photographs to 
be posted after its purpose was explained. A total of 22 new items 
were obtained for our two websites during the event. The largest 
number of contributions came from one family who brought 15 
original historical land leases, 9 of which have been included on our 
website. 

This revealed a flaw in our planning. While volume of 
contributions had been considered, it had seemed very unlikely that 



The Caprons of Paris 219 



there would be a surfeit of relevant materials for inclusion on our 
website. Consequently, our planning had focused on limiting the 
number of contributions per person. Fortunately, contributors were 
evenly spread across the day; in the future, we will ensure that all 
relevant material can be digitized in the event that the number of 
contributors exceeds capacity. 

The Digitization Days event proved to be a success. The most 
valuable outcome was a new contact who owns a large untapped 
collection of original historical documents. 

Our project's launch event was held on December 6, 2009. We 
invited members of the community to attend and see an introduction 
to our two digital collections websites. As entertainment, we provided 
a slide show with photographs from the collection. In addition, the 
president of the historical society we partnered with came in full 19th 
century costume in character as "Hiram Capron." We sent Victorian- 
styled invitations to a wide variety of recipients, and attracted visits 
from our Member of Provincial Parliament and the County's mayor. 

Conclusion 

While the two Digital Collections sites have now officially 
launched, development is very much ongoing. The materials available 
have acted as catalysts for community interest and involvement. 
Based on the current interest the project has generated, the Library 
has continued its work with the Paris Museum and Historical Society 
for additional projects and partnered with new organizations to 
document the history of other communities in the County. By 
digitizing the wide body of materials described, and opening doors 
within the community, the Capron project has helped lay the 
foundation for a digitization program which can document the history 
of all of the County's communities. 

References 

Adobe. (2009). Digital Negative (DNG). Retrieved Dec 8, 2009, from 
http : / / www. adobe .com/products/ dng/ 



220 The Caprons of Paris 

Atiz. (n.d.). Atiz BookSnap: It's a book ripper. Retrieved from 
http : / / www. atiz. com /brochure /booksnap . pdf 

Baker, T., et al. (2009). Collaborative history - Creating (and 
fostering) a wiki community. In J. Trant and D. Bearman (eds). 
Museums and the Web 2009: Proceedings. Toronto: Archives & 
Museum Informatics. Retrieved from 

http://www.archimuse.com/mw2009/papers/baker/baker.html 

Butler, Richard. (2009, December 16). Canon PowerShot G11 Review. 
Digital Photography Review. Retrieved Dec 16, 2009, from 
http : / / www. dpreview. com / reviews / canongn/ 

Canadian Museum of Civilization Corporation. (2007). Digitization 
Standards for the CMCC: Scan and Artifact Photography. 
Retrieved May 17, 2009, from 

http://www.chin.gc.ca/ATutor/bounce.php?course=29 

Canon. (2008). PowerShot G10 Camera User Guide. Japan: Canon. 

Clements, Maureen. (2009). The secret of Google's book scanning 
machine revealed. As a Matter of Fact Blog: NPR. Retrieved from 
http : / / www. npr.org/blogs/library/2009/04/ 
the_granting_of_patent_75o8978.html?sc=fb&cc=fp 

DxO Labs. (n.d.). Compare cameras. DxO Mark. Retrieved Dec 11, 
2009, from http://www.dxomark.com/index.php/eng/Image- 
Quality-Database/ Compare-cameras/ (appareili)/247| 0/ 
(apparent) /334 1 0/ (apparent)/ 3 19 1 0/ (onglet) /o/ (brand) / Canon/ ( 
brand2)/ Canon/ (brand3)/ Canon 

Epson. (2009). Epson GT-20000. Retrieved Nov 3, 2009, from 
http://www.epson.com/cgi-bin/Store/consumer/consDetail.jsp? 
BV_UseBVCookie=yes&oid=63075503 

Epson. (2009). Perfection V500 Photo Color Scanner. Retrieved Nov 
3, 2009, from http://www.epson.eom/cmc_upload/o/ooo/142/ 
325 /V500_InfoSheet.pdf 

Hess, Richard L. (2009). Re: NEF images. Aug 13, 2009. ARCAN-L. 
http://www.mailman.srv.ualberta.ca/mailman/private/arcan- 
l/2009-August/oo6i36.html 



The Caprons of Paris 221 



Holden, Frederic A. (1859). Genealogy oftheDescendents ofBanfield 
Capron. Retrieved Dec 8, 2009, from 
http://books.google.ca/books?id=qLHTscb_XEwC 

Holden, Frederic A. (1859). Genealogy oftheDescendents ofBanfield 
Capron. Boston: Geo. C. Rand & Avery. Paris Museum and 
Historical Society, 2005.0044.02. 

King Township Public Library. (2008). Timeless king online. 
Retrieved Dec 14, 2009, from http://www.king- 
library.on.ca/heritage.php 

Knowledge Ontario. (2009). VITA [3.3] user manual (Version 1.0). 
Retrieved from http://our-ontario-contributor-network.near- 
time.net/files/vita3-3_usermanual.pdf 

Qua, George. (2009, October 12). Qua family name. Ulster ancestry 
genealogy & ancestry forums. Retrieved Dec 10, 2009, from 
http : / / www. ulster ancestry . com/ forums/ viewtopic. php ?t= 28 o 1 

Ristech. (2009). Kirtas APT BookScan 1200. Ristech. Retrieved Nov 
3, 2009, from http://www.ristech.ca/kirtas-apt-bookscan- 
1200.html 

Semantic MediaWiki. (2009). Help:Inline queries. Retrieved Dec 1, 
2009, from http://semantic-mediawiki.org/wiki/ 
Help : Inline_queries 

Statistics Canada. (2006). 2006 community profiles - Brant. 
Retrieved Dec 14, 2009, from http://wwwi2.statcan.ca/census- 
recensement/20o6/dp-pd/prof/92-59i/details/page.cfm? 
Lang=E&Geoi=CSD&Codei=3529005&Geo2=PR&Code2=35&Dat 
a=Count&SearchText=brant&SearchType=Begins&SearchPR=oi& 
Bi=All&Custom= 

Torborg, Wayne. (2008). Manuscript digitization at the Hill Museum 
& Manuscript Library: Building on the legacy of microfilm. 
Microform & Imaging Review. 37, 1, pp. 17-27, ISSN (Print) 0949- 
5770, DOI: I0.i5i5/mfir.2008.002, Winter 2008 

U.S. National Archives and Records Administration. (2004). 
Technical guidelines for digitizing archival materials for 
electronic access: Creation of production master files - raster 



222 The Caprons of Paris 



images. Retrieved from http://www.archives.gov/preservation/ 
technical/ guidelines.html 

Van Ballegooie, Marlene, and Duff, Wendy M. (2001). RAD revealed: 
A basic primer to the rules for archival description. Ottawa: 
Canadian Council of Archives. 

Wan, Don. (2008). Canon PowerShot G10 review. Digital 
Photography Review. Retrieved Dec 2, 2009, from 
http : / / www. dpr eview. com/ reviews / CanonGio/ 

Wikia. (2009). Help:Rich text editor. Wikia. Retrieved Nov 27, 2009, 
from http://help.wikia.eom/wiki/Help:Rich_text_editor 

Zurbu. (2009). Local history and development forums on Zurbu. 
Retrieved Dec 4, 2009, from http://zurbu.net/ 



The Mass. Memories Road Show: A State- 
Wide Scanning Project 



Joanne Riley and Heather Cole 
(University of Massachusetts Boston) 



Abstract 

Running a state-wide digital history project on a shoestring budget 
and staffed primarily by volunteers is not only possible, but brings 
immeasurable rewards for the contributors, volunteers, organizers 
and staff while gathering priceless documentation of their communal 
heritage. The Mass. Memories Road Show (http://www. 
MassMemories.net) is a public scanning project based at the 
University of Massachusetts Boston which partners with local 
communities to digitize family photographs and stories at public 
events with the goal of creating a digital portrait of all the 351 cities 
and towns in the Commonwealth. This article describes how the 
project works to ensure broad participation in the planning and 
execution of the project, as well as a detailed description of the 
logistics of a Road Show event, which could be replicated in other 
communities. 

Keywords: Archives, Community history, Community, Digital 
collection, Digitization, Family, History, Libraries, Local history, 
Massachusetts, Memories, Multi-cultural, Partnership, Photos, Place- 
based, Public history, Public, Scanning, State studies. 



223 



224 The Mass. Memories Road Show 



Introduction 

The Mass. Memories Road Show (MMRS) is an ongoing, on-the-spot 
public scanning project in the Commonwealth of Massachusetts, 
sponsored by the Joseph P. Healey Library at the University of 
Massachusetts Boston and Mass Humanities, the state humanities 
council. The MMRS documents Massachusetts people, places and 
events through the contributions of individuals who bring their photos 
and stories to be digitized at public events throughout the state. Over 
the next few years, the project will partner with hundreds of local 
organizations to visit each of the 351 communities in Massachusetts, 
gradually building up a self-portrait of the Commonwealth through 
the contributions of its residents. The MMRS is online at 
http://www.MassMemories.net. 

Project Background 

The MMRS grew out of the place-based education initiatives of the 
University of Massachusetts Boston's "Massachusetts Studies Project" 
(MSP), which provides resources for Massachusetts teachers and 
students in the areas of local history, culture and environmental 
studies. A series of casual brainstorming sessions with librarians, MSP 
board members, teachers and local historians coalesced in a vision of 
a public history project inspired by elements of PBS's Antiques 
Roadshow (people bringing their personal treasures to a local event 
for professional perusal) and the Library of Congress' American 
Memory Project (a library organizing digitized images from a common 
heritage to be shared on the World Wide Web, see The Library of 
Congress, n.d.) The subsequent development of the MMRS project 
was guided by the work of Daniel Cohen and Roy Rosenzweig (2005), 
Stuart Lee and Kate Lindsay (2009) and by the writings of Robert 
Putnam and others about the nature of civic engagement. We have 
borrowed elements from a number of superb online digital history 
projects, among them the Maine Memory Network (Maine Historical 
Society, 2009), for its richly productive collaborations with partnering 
organizations across the state, The Organic City (The Organic City, 
n.d.) for its place-based communal storytelling approach, the Coney 



The Mass. Memories Road Show 225 



Island History Project (Coney Island History Project, n.d.) for its 
effective application of social web tools, the Charlestown Digital Story 
Project (UMBC Digital Story, n.d.) at UMBC for its engaging 
multimedia oral histories created through student-elder 
collaborations, the Worthington Memory Project (Worthington Public 
Library, 2002) for its transparent application of best practices in 
indexing, and Orlando Memory (Orlando Memory, n.d.) for its 
solicitation of public contributions to a community history project. 

The MMRS project was originally designed to meet two goals: 
collecting digital surrogates and personal annotations of locally held 
primary sources that document people, places and events in 
Massachusetts; and developing a searchable online repository of 
sources that could be used for educational purposes at all levels. As 
the project developed, we discovered that it met another important 
need that has been incorporated as a key goal of the project: 
community building. Road Shows have turned out to be deeply 
engaging community events that connect people within the 
community to each other and to others throughout the state, and have 
proven to be meaningful in lasting ways to the people who contribute 
and to those who volunteer. At its best, the MMRS seems to generate 
both of the types of productive social relationships that Robert 
Putnam describes in Bowling Alone: the Collapse and Revival of 
American Community (Putnam, 2000); that is to say, bonding social 
capital (holding together people who see themselves as being similar 
in social identity) and bridging social capital (bringing together people 
who consider themselves to be "unalike" in some aspect of social 
identity.) In the words of a Road Show volunteer and contributor: 

"[The Road Show] brings the community together... [It] 
acknowledges to people that they are a part of the history- 
making... It brings people to understand and respect different 
cultures in their town through the old and new pictures... I think 
it is one of the great community programs to encourage the 
whole family to be involved." (MMRS Letters of Support, 2008). 

Since its launch in the fall of 2004, the MMRS project has 
organized Road Show events across the state, in the process digitizing 



226 The Mass. Memories Road Show 



thousands of photos and stories, collaborating with dozens of 
community organizations and generating the practical lessons we will 
share in this essay. 

Guiding Principles 

The following key principles have guided the development of the 
MMRS since its inception: 

First, we believe that everyone who lives or lived, works or 
worked in a given community is equally representative of that 
community. Thus, the Road Show's on-the-ground events and the 
resulting digital collection are intended to reflect as realistic a picture 
as possible of the community, based on demographics past and 
present. For example, for a community that was predominantly Irish 
in the early 20th century and is now largely Vietnamese, the event 
attendees and the photos and stories in the digital collection should 
reflect both of those cultures. 

A second guiding principle of the MMRS project is the idea that 
we are not creating a comprehensive archive of Massachusetts' (or a 
given community's) history; rather, we are creating a self-portrait of 
the state based on the photographic artifacts that are held in 
individuals' personal photo albums and scrapbooks. Thus, the photos 
that individuals bring to the event reflect their choices of how they 
want to represent themselves and their families in the project archive, 
as opposed to meeting a predetermined collecting standard. This leads 
naturally to the project protocol that contributors caption and 
describe their photos in the first person, rather than in impersonal 
archival language. For instance, one Road Show image's caption and 
description read: 

"The Reading Theater Circa 1924: This is the second theater in 
Reading built by my great uncle, Ed Turnbull. He also built the 
first in 1913. My father, Arthur Kelley, was a popcorn boy. We 
moved to Reading because my father loved it after spending so 
much time here." (Mass. Memories Road Show, 2010; Image ID 
41.106.1). 



The Mass. Memories Road Show 227 



Another guiding principal is that the Road Show events 
themselves play a vital role in creating an understanding of communal 
history. Particularly in large, diverse cities, individuals and 
organizations may not be aware of the larger historical and 
contemporary context of their community. Yet all of the people 
arriving at their local Road Show with photos in hand have in 
common their connection to that physical location and its history, 
regardless of their age, ethnic, economic or religious background; all 
"belong" simply by virtue of their involvement (present or past) in this 
place. At the Road Show events, each person's connection to the 
community is formally acknowledged through their presence there. 
The events help build connections between contributors and 
contributing organizations and the digital collection serves to mirror a 
community back to itself. This principle of inclusion informs all 
planning processes and influenced our early decision to arrange the 
public events so that all participants could observe and share in the 
contributions of others. For example, we project all photos on a wall 
screen as they are being scanned, and arrange seating so that 
contributors can listen to others' stories as they are being recorded. 

A final, later-adopted guiding principal of the MMRS is that the 
contributor is as worthy of documentation as the contribution. We 
hope that our digital collection will serve as a resource of primary 
source material well into the future. As such, we came to understand 
that the participation of Road Show contributors and volunteers was 
also worthy of documentation. Each person who takes part in a Road 
Show has shown a commitment to their community history that we 
realized should be captured and preserved. Thus, we include 
"Keepsake Photos" (described below) and staff / volunteer group 
photos in the dataset for each Road Show. 

These principles formed the foundation for the development of 
the MMRS project and are, we feel, responsible for the extraordinary 
atmosphere of cooperation and community that participants have 
mentioned in describing the Road Shows: 

"The Road Show connected people of all ages to the past, 
resulting in a deeper understanding of the present. [The] Road 



228 The Mass. Memories Road Show 



Show modeled for youth how to work together to share the 
history of everyone's history. Through involvement in the Road 
Show, a torch was passed to [our] youth, the torch of stewardship 
of their community's history." - a Road Show 
organizer/ contributor (Letters of Support, 2008). 

"Many participants stayed and shared their photographs with 
those sitting next to them, watched others being videotaped, or 
watched the images being projected onto the overhead screen. 
The interactive nature of the event allowed participants to share 
stories and learn about other communities in their city." - a Road 
Show organizer (MMRS Post-Event Participant Survey, 2009). 

The Massachusetts Model 

From the beginning, we aimed to create a model project that could be 
replicated beyond Massachusetts. In the second year of the project, we 
consulted with archival, legal and information technology specialists 
to ensure that the database structure and metadata collection forms 
met accepted standards and best practices for digital history projects 
and that the online resources would available for educational use. 
These specialists' recommendations were compiled in The Mass. 
Memories Road Show Handbook: Procedures and Protocols for a 
Public Scanning Project (University of Massachusetts Boston, 2006), 
available on the project website. They guided us in creating an open- 
source database system that stores the information from the 
Contributor and Photo forms in fields that are compatible with the 
Dublin Core metadata set, and in ensuring that the images collected 
may be made available online on a non-exclusive basis for non- 
commercial uses. 

Road Show Logistics 

Funding 

The early years of the MMRS project's development were partially 
supported by grants from the Massachusetts Historical Records 
Advisory Board, the Mass. Foundation for the Humanities, and a 



The Mass. Memories Road Show 229 



Professional Development grant program at the University of 
Massachusetts Boston. These grants, totaling approximately $30,000 
over three years, covered some initial equipment, student work, and 
expenses for pilot Road Shows. The majority of the project's support 
in its early years was in the form of extensive contributions of time 
and equipment from the project director, volunteers and community 
collaborators. Thanks to the track record built up through the early 
Road Shows, the project reached a turning point in 2006, when it was 
brought under the umbrella of the University of Massachusetts 
Boston's Joseph P. Healey Library as a community outreach initiative, 
a step which allowed for the hiring of a public historian as a project 
manager (.5 FTE) to focus on growth, stabilization and outreach. 

In 2008, the state humanities council, Mass Humanities joined 
with UMass Boston's Healey Library to co-sponsor a "mini-grant" 
program to expand the Road Show. The program accepts applications 
from interested communities who are required to assemble a local 
planning team and outline how they plan to achieve a realistic 
community portrait through their outreach and marketing. If accepted 
for funding, the planning team handles local outreach and provides 
volunteer staff, translators, non-technical equipment and 
refreshments on the day of the event, while the grant funds 
professional videography, stipends for event staff and post-production 
of images, data and videos. The Healey Library provides technical 
equipment and in-kind staffing for preparation, volunteer training, 
website application support, consultation with the University 
Archivist, running the Road Show event and follow-up. 

Planning 

Planning for a Road Show begins as much as a year in advance of 
the event. Because our goal is to make these events and the materials 
collected a realistic self-portrait of the community, we involve as many 
local organizations as possible in the planning process: public 
libraries, historical societies, public schools, cultural and ethnic 
organizations, youth groups, historic preservation organizations, 
business people, government officials, genealogists, senior citizens 
and others. 



230 The Mass. Memories Road Show 



Ensuring participation by those who are sometimes under- 
represented in local history projects (young people, non-English- 
speakers, new arrivals and other minority members of the 
community) is most dependably accomplished when the local 
planning team is comprised of persons representing the same range of 
backgrounds as the community at large. While this representation is 
critical for the Road Show's success, it is equally important that local 
planning team members share a genuine interest in cultural heritage, 
archives, genealogy, family history, community history and 
community building. Too often, "diversity" in planning committees is 
arrived at by requesting the participation of municipal or 
organizational representatives who are routinely tapped to represent 
their communities in all sorts of projects, regardless of their personal 
interests. We have learned that putting extensive effort into finding 
true kindred spirits across a community's various subgroups - those 
whose eyes light up at the mention of "old pictures" or "family stories" 
or "local history" - can pay huge dividends in the overall ease of 
encouraging broad participation in the eventual Road Show. 

The local planning team is responsible for outreach to the 
community to recruit contributors to the event. We have found that 
the best-attended Road Shows were invariably the result of person-to- 
person outreach by the local organizing team which, while time- 
consuming, was much more effective than general publicity or 
postings around town. 

One community member described her participation in a local 
planning team as valuable both personally and professionally: 

"The Mass. Memories Road Show is a vehicle not only to 
showcase communities but helps to bring them together. I was 
able to observe how members of various and diverse institutions 
came together to put on a community event. I met leaders of the 
Asian community, historic/ cultural institutions and veteran 
neighborhood organizations" - MMRS local organizing 
committee member (MMRS Letters of Support, 2008). 



The Mass. Memories Road Show 231 



Staffing 

The permanent MMRS staff consists of a part-time project 
manager who oversees all aspects of the project, and the MSP director 
who serves in an administrative and advisory capacity. Both positions 
are located within the Digital Library Services Department of the 
Healey Library at the University of Massachusetts Boston. 

In addition to the permanent MMRS staff, each Road Show event 
is staffed by 15-20 people: a combination of experienced "Roadies" 
(people who previously volunteered at a Road Show in their own 
community) and local volunteers. Roadies receive a small stipend (ca. 
$50) for each subsequent Road Show they work; in other words, staff 
are not paid to work in their own community but are compensated to 
bring their experience and expertise to another Road Show in the 
state. Local volunteers are recruited by the local planning team and 
need not have any professional experience in archives, local history or 
digital projects. We have been pleased to have skilled amateur 
photographers, genealogists, librarians, teachers, historians and 
computer buffs among our volunteers and Roadies. 

One of the many unexpected outcomes of this project has been 
the camaraderie built up among Roadies across Road Shows, and the 
cross-fertilization that occurs through the activities of people from 
different communities traveling to help others document their 
community heritage. Some Roadies have been with the project for 
several years, and the number grows with each event, since all 
contributors and volunteers are routinely invited to help staff 
subsequent Road Shows. 

Staffing an event is apparently equally rewarding for the Roadies, 
who often speak of feeling privileged to bear witness to the stories 
shared by event contributors: 

"Once the event began, I was touched in many ways. Some 
stories brought a tear to my eye, like one woman's description of 
how her brother was a prisoner of war for six months in World 
War II, and the hurt and pain she felt, but the comfort his 
picture brought her... Working on the project is by far the most 
rewarding venture I have been involved in so far." - MMRS 



232 The Mass. Memories Road Show 



volunteer, Roadie and videographer (MMRS Letters of Support, 
2008). 

Training 

All of the Road Show functions aside from videography - i.e. 
photo scanning, metadata collection, still photography and customer 
service - are performed by amateur Roadies and volunteers who are 
trained just prior to the event. The MMRS staff provides written 
instructions and conducts the training for all Roadies and volunteers. 

In the past, volunteer training had been scheduled two days 
before the Road Show, which allowed time to run the training along 
with an interim day to fix any problems that might be identified 
during the training session. However, it proved difficult to assemble 
all of the volunteers in advance of the event, so subsequent Road 
Shows have incorporated an hour-long training immediately before 
the event, when all volunteers for the day are present. While this 
worked reasonably well, we are developing a series of self-paced 
multimedia training materials for all volunteers to work through well 
in advance of the event. This will improve the efficacy of the training 
process and will also bolster the confidence of inexperienced local 
volunteers as the day of the Road Show approaches. 

AT A ROAD SHOW 

Each Road Show is held in a public space, most often a local public 
library, historical society or community center. Most Road Shows are 
scheduled to last about four hours, with an additional two hours in 
advance for setup and an hour for breakdown and cleanup. Each Road 
Show consists of "stations", which are tables devoted to specific Road 
Show functions. Road Show participants move from one station to 
another, providing information, having their photos digitized and 
their stories recorded. 

Contributors are invited to drop in at any time during the event. 
We have found it works best to have a loose and casual atmosphere. If 
a contributor arrives under great time pressure, we squire them 
quickly through the process. But for the most part, people arrive at the 



The Mass. Memories Road Show 233 



event, go through the process, and then linger to listen and watch as 
others go through the process. Even when the crowds have been 
heavy, there has never been any impatience but only excitement and 
goodwill. Said one contributor: "7 could have spent two hours with all 
my photos and stories. I really had a good time" (MMRS Post-Event 
Participant Survey, 2009). 

Below are details about each of the seven Road Show event 
stations, offered in hopes that those who choose to undertake a similar 
endeavor may benefit from lessons learned in this project's early 
years. 

Welcome Station 

The Welcome Station is the first stop for all contributors to the 
Road Show. Locating the Welcome Station in the corridor or 
anteroom outside the event space helps manage crowds and can also 
serve as a visual invitation to the event for passersby. The station is 
staffed by two "Greeters", generally one experienced Roadie and one 
local community member. As contributors approach the table, the 
Greeters welcome them to the event, briefly explain the station layout 
and give them a numbered nametag and a Registration Form to 
complete. The number on the contributor's nametag and form is used 
to match each contributor to his/her photos, stories and metadata. 
Greeters accept and review the Registration Form, and then invite the 
contributors to proceed to the Info Station to complete paperwork on 
their photos. If any of the contributors need assistance with 
translation, mobility or other challenges, then a floating volunteer is 
summoned to assist them as they proceed from station to station. 

Equipment: 4-6' table and chairs; pre-numbered stick-on 
nametags; registration forms; clipboards; pens (all forms must be 
signed in pen.) 

Info Station 

At the Info Station, contributors complete paperwork on each of 
the photos they are contributing to the MMRS digital collection. The 
Info Station is staffed by three to five "Reviewers." Their role is to 
assist contributors in selecting a few photos that represent themselves, 



234 The Mass. Memories Road Show 



their family and/or their community, and assist them in completing a 
Photo Form for each photo. Reviewers encourage contributors to 
select photos that best fit one or more of the following categories: an 
original photo (rather than a printout or photocopy), a photo where 
they know some identifying details (where taken, when, people 
photographed, etc.), a photo that has particular meaning to the 
contributor, a photo that is unusual in some way, or a photo that has 
people in it (rather than a scenic shot). 

Contributors are encouraged to complete the Photo Form as best 
they can, using a first-person approach in providing the "who, what, 
when, where" of the photo. The Reviewer then reviews the Photo 
Form and may prompt the contributor to clarify or add more 
information (last names, additional details, etc.) Reviewers can add 
more information derived from conversation on the back of the form, 
if necessary. Once each Photo Form is complete, the Reviewer assigns 
a sequential number to each of the contributor's photos, writing that 
number on the relevant Photo Form. Finally, the Reviewer invites the 
contributor to take their photos and Photo Forms to the Scanning 
Station, accompanying the person there if necessary. If there is a 
waiting line at the Scanning Station, the contributor may visit any of 
the other stations in any order. 

Notes: In our experience, reviewers need to actively encourage 
contributors to use first person ("I, Me, My, Our...") in photo captions 
and descriptions. Frequently contributors will bring in an entire 
album of photographs and it can take some time to help them sift 
through them to select two or three to share with the project. 
Contributors frequently do not fill out the Photo Forms completely at 
first; it often takes additional coaching or asking questions to 
encourage the contributors to share more detailed information about 
their photos. 

Equipment: Two 6-8' tables 10-12 chairs; Photo Forms; 
Clipboards; Pens 

Scanning Station 

At the Scanning Station, the contributor's photographs are 
digitized on a flatbed scanner. This station is staffed by two 



The Mass. Memories Road Show 235 



volunteers/ Roadies: a "Scanner" and a "Reviewer." The Reviewer 
collects the Photo Forms and performs another quality control check 
for completeness, legibility and correctness of the ID number. The 
Scanner scans each photo at 300 DPI and saves as a TIF file named 
using the format RoadShow#.Nametag#.Photo#.tif. This filename is 
also recorded on the Photo Form. The scanning process is projected 
live on a wall screen for all event visitors to see, which enhances the 
spirit and enjoyment of the event. After the photos are scanned, they 
are immediately returned to the contributor, while all of the 
paperwork is collected by the Reviewer and stored in a large envelope 
at each station. 

If the photo is too big or unwieldy to fit on the scanner, or if, as 
sometimes happens, the contributor brings a three-dimensional object 
(we have seen aprons, clocks, weavings, jewelry and more) those 
contributors are sent to the Keepsake Photo Station where the digital 
camera is used to photograph the oversized object. 

Notes: Each Scanning Station can handle between 70 and 100 
photos within the event time. For most Road Shows to date, this has 
meant setting up two Scanning Stations. 

Projecting the scanning process onto a wall screen not only helps 
to build community, but also is an engaging way for those waiting 
their turn at one of the stations to pass the time. There have been a 
number of exciting moments in Road Show where someone recognizes 
a person or place that flashes on the screen and adds their story to the 
mix. 

Equipment: laptop computer; scanner (capable of scanning 
letter-sized documents and photographs as tiffs at 300 DPI) and 
necessary drivers to connect scanner to laptop; LCD projector; 
portable projection screen; paperclips; blank CDs for backup; two 6-8' 
tables and chairs; large envelopes to hold collected photo and 
contributor forms and backup CD. 

Keepsake Photo Station 

The Keepsake Photo Station is staffed by a "Photographer" and a 
"Paperwork Manager." The Photographer snaps a photo of the 



236 The Mass. Memories Road Show 



contributor holding one (or more) of the contributed photos, from a 
distance of about four feet, and framing from the waist up. The 
Photographer then prints out the photo as a souvenir for the 
contributor to take home with them, while the digital version is saved 
for later incorporation into the online database. The Paperwork 
Manager is responsible for completing the Keepsake Photo Log with 
the name, nametag number and other relevant information which is 
transferred to the image file of the keepsake photo after the event. 

Notes: The Keepsake Photo Station also serves as the station for 
photographing images and objects that are too large to fit on the 
scanner. Because the printing process can be slow, contributors are 
invited to visit other stations while their keepsake photo is being 
generated. Printed photos are posted onto a bulletin board for later 
pickup. 

Equipment: digital still camera; tripod; keepsake photo log; rich 
colored backdrop; photo printer with paper and ink (capable of 
printing 4x6" prints from digital camera onsite); bulletin board and 
thumbtacks; table for printer; chairs 

Video Station 

At the Video Station, contributors are asked to share a 3-5 minute 
story about the photo(s) that they contributed to the project. At the 
Video Station there is a Videographer and an Interviewer, although 
these can be the same person, depending on the experience of the 
staff. The Interviewer begins by asking the contributor to state their 
name and nametag number on camera for data quality control 
purposes, and then asks the contributor to share a three to five minute 
story about one or more of the photos that they brought with them. 
Clarifying questions may follow, in a casual, conversational style. 

Notes: While trained amateurs can achieve high quality scanning 
results, videography is still best left to professionals, or at least those 
with extensive experience in videotaping in public settings under time 
and quality pressures. Public access television stations in each 
community have provided experienced staff for this important 
function. 



The Mass. Memories Road Show 237 



Equipment: digital video camera; appropriate lighting, lavaliere 
microphone and other equipment for recording interviews; 
photogenic chair with armrests for interviewees to sit in; backdrop, 
preferably one reflective of the community. 

Preservation Station 

The Preservation Station includes a display and handouts on how 
to care for family photos and documents, both contributed by 
volunteers from the New England Archivists. Contributors can stop by 
this table to peruse the displays, pick up information and ask 
questions of the archivists. Road Show volunteers direct contributors 
to the Preservation Station if they notice that a contributor's artifacts 
are delicate or in poor condition, or in response to questions from 
contributors about preservation, valuation or potential repositories for 
their artifacts. 

Notes: Despite the Preservations Station's alluring display of 
"Archival Evils" showing the dramatic effects that mold, moisture, 
heat, glue, etc can have on precious family artifacts, contributors have 
tended to pass by this station, perhaps because it is not directly 
related to the digitization processes. To mitigate this, we now place 
the Preservation Station closer to the scanning action, and encourage 
the archivists to wander the room and offer advice as moved to do so 
by what they see among the contributions. 

Equipment: 6-8' table and chairs; display materials and 
handouts about photo preservation, preferably supplied by 
professional archivists 

Local History Station 

The Local History Station provides consultation and resources on 
local history. This station is usually staffed by representatives from the 
local historical society or reference librarians who are broadly 
knowledgeable about the history of the area and its residents over 
time. The Local History Station includes a display on local history and 
local history resources (books, pamphlets, maps, etc.) Local 
historians/librarians also keep an eye out for any photos that they 
may want to solicit as a donation to their archives. 



238 The Mass. Memories Road Show 



Notes: The volunteers at the Local History Station are able to 
answer local questions, and identify persons, places and dates in 
contributors' photos. These are also the people who will be carrying on 
aspects of the project into the future. Usually the local groups who 
staff this station come away with new member signups to their 
organizations. 

Equipment: 6-8' table and chairs; brief written history of the 
town in pamphlet form; other relevant local history information 
(books, maps, etc.) 

Post-Event Processing 

Based on our own experience and the recommendations of 
professional consultants, we developed a routine for managing the 
wealth of information gathered at each Road Show. At the end of each 
event, we back up the digital images onto a portable hard-drive and 
make photocopies of all the paperwork. Within a few weeks of each 
event, MSP staff members perform the data-entry and image 
processing necessary to include the images and metadata in the online 
database. Once data entry is complete, we send digital copies of all 
materials to each partnering organization for that event. We also send 
out a brief survey to participants, send contributor contact 
information to all partners, and finally, notify all participants when 
the photos and videos have been mounted on the project website at 
www.MassMemories.net. 

351 and Beyond 

In its first five years of operation, the MMRS has visited only a 
fraction of the 351 cities and towns across Massachusetts and has 
collected just a portion of the millions of photos and stories that still 
lie in the basements, attics, scrapbooks and photo albums of its 
residents. We have a long way to go and, like many other projects, 
continue to seek ways to reach a broader audience, ensure long-term 
sustainability, and reduce operational costs. To that end, we are in the 
process of converting the project's current database and website into a 
web-based data-management system to streamline data- and image- 
processing procedures. 



The Mass. Memories Road Show 239 



At Road Shows we have seen high school students and retirees 
share stories about how their common childhood neighborhood has 
changed over time. We have heard former residents from opposite 
sides of what had been a racially-segregated housing development 
recall overlapping events from their childhoods. Parents have come to 
Road Shows with their children to share photos of their immigrant 
ancestors. After the events, we see communities continuing the 
connections and conversations that were started at Road Show events 
through local history exhibits, extended documentaries about local 
residents, and new partnerships among local organizations. The 
ultimate lesson learned? Running a state-wide digital history project 
on a shoestring budget and staffed primarily by volunteers is not only 
possible, but brings immeasurable rewards for the contributors, 
volunteers, organizers and staff while gathering priceless 
documentation of their communal heritage. 

In closing, here are words from some of those who have 
participated in the Mass. Memories Road Show to date: 

"The students said they learned so much from the families and 
individuals who showed up to share photos and tell stories of the 
old days... Students began to realize that they will someday be the 
elders, the caretakers of their city's, their family's, their culture's 
history." - Middle school teacher whose students participated in a 
Road Show as contributors and volunteers 

"I met many new people and heard many interesting stories 
about the history of the community in which I live that I would 
have otherwise never heard. I have left the experience of my 
involvement with the Mass. Memories Road Show with even 
more interest in my community and the tapestry of people that 
populate it." - Road Show volunteer/contributor. 

"I have only worked in three Road Shows and can't wait for 
more!" - 8th grade student who worked as a Roadie during his 
summer break. 

"We heard the story of the lamplighter in North Quincy, learned 
that a Quincy resident was Jimmy Dolittle's wingman, and that 
people still remember the smell of fresh bread at the Sumner 



240 The Mass. Memories Road Show 



Bakery. These footnotes don't appear in standard histories, but 
now they are on record for all to share." - Local history librarian 
and Road Show organizer. 

"Having spent my entire career in the field of public history, I 
cannot think of a project that contributes more to a broader 
public reflection and appreciation of the past. The Road Show has 
developed a creative, innovative and, perhaps most importantly, 
an effective approach to engaging a community in documenting 
its own history." - State humanities council representative and 
Road Show attendee 

The authors would like to thank the following supporters who 
helped shape the Mass. Memories Road Show in its earliest days: Ron 
Adams, Phil Byrnes, Celeste Finison, George Hart, April Hagins 
Johnson, Elizabeth Clancy Lerner, Cagen Luse, Marisa Luce, Mary 
McCarthy, Paul McCarthy, Elisabeth McGregor, Hoa Mai Nguyen, 
Daniel Ortiz-Zapata, Frances Pollitt, Frank Poon, Barbara (Bobby) 
Robinson, Ellen Rothman, Sequoia Stenlund, Emily Sweeney, Earl 
Taylor, Rebecca Withers. 

References: 

Cohen, Daniel J. and Roy Rosenzweig. (2005), Digital history: A 
guide to gathering, preserving and presenting the past on the 
web. University of Pennsylvania Press. Retrieved March 23, 2010 
from http:/ / chnm.gmu.edu/ digitalhistory. 

Coney Island History Project, (n.d.). Retrieved March 23, 2010 from 
http://www.coneyislandhistory.org/ 

Lee, Stuart D. and Kate Lindsay. (2009). If you build It, they will 
scan: Oxford University's exploration of community collections. In 
EDUCAUSE Quarterly, vol 32, 2009. Retrieved March 23, 2010 
from http://www.educause.edu/EDUCAUSE+Quarterly/ 
EDUCAUSEQuarterlyMagazineVolum/IfYouBuildltTheyWillScan 
Oxford/i74547#TB_inline?height=400&width=630&inlineId=side 
bari&modal=false 



The Mass. Memories Road Show 241 



Maine Historical Society. (2009). The Maine Memory Network, a 
project of the Maine Historical Society. Retrieved March 23, 2010 
from http://www.MaineMemory.net. 

Mass. Memories Road Show. (2010). Mass. Memories Road Show - 
Your family's place in Massachusetts History. Retrieved March 23, 
2010 from http://www.massmemories.net. 

MMRS Letters of Support. (2008). Letters of support submitted for 
MMRS nomination for NCPH Public History Award. Unpublished 
documents. (January, 2008). 

MMRS Post-Event Participant Survey. (2009). Unpublished data. 

Orlando Memory (n.d.). Homepage. Retrieved March 23, 2010 from 
http://dc.ocls.info/. 

Putnam, Robert D. (2000). Bowling alone: the collapse and revival of 
American community. Retrieved March 23, 2010 from 
http://www.bowlingalone.com/ 

The Library of Congress, (n.d.). The Library of Congress: American 
memory. Retrieved March 23, 2010 from http://memory.loc.gov/ 
ammem/index.html. 

The Organic City. (n.d.). Homepage. Retrieved March 23, 2010 from 
http : / / www. theorganiccity. com/ wor dpr ess / . 

UMBC Digital Story, (n.d.). Digital stories from Charlestown. 
Retrieved March 23, 2010 from http://www.umbc.edu/oit/ 
newmedia/ studio/digitalstories/ctds.php 

University of Massachusetts Boston. (2006). The Mass. Memories 
Roadshow handbook: Procedures and protocols for a public 
scanning project. Retrieved March 23, 2010 from http://www. 
msp.umb.edu/MassMemories/handbook/MMRSHandbook.pdf. 

Worthington Public Library. (2002). Worthington memory. 
http : / / www. worthingtonmemory. org/index. cfm. 



Picturing the Museum: Education and 
Exhibition at the American Museum of 
Natural History 

Kelli Anderson, Barbara Mathe, Eric Muzzy, Stacy Schiff 
(American Museum of Natural History Research Library) 



Abstract 

The Research Library of the American Museum of Natural History 
received funding from the Metropolitan New York Library Council in 
2007 to produce a web exhibit of 989 historic images. Picturing the 
Museum: Education and Exhibition at the American Museum of 
Natural History served as the prototype for a comprehensive 
database for the Research Library's extensive Photographic Collection. 
In the larger context of the history, development, and use of the 
Library's Photographic Collection, this article describes the project's 
conception through a self-published book produced for the Trustees' 
Library Committee, the funded project and the ongoing development 
of the larger database. One of the internal goals and results of the 
Picturing the Museum project was to analyze, codify, and document 
local practice, policies and workflow for more efficient delivery of 
images to the web. 

Keywords: Cultural metadata, Digital imaging, Digitization, 
Exhibition, Libraries, Museum, Museum education, New York, 
Photograph, Photographic collection. 



242 



Picturing the Museum 243 



The American Museum of Natural History 
Photographic Collection 

Founded in 1869, the American Museum of Natural History (AMNH) 
is one of the nation's preeminent institutions for scientific research 
and public education. Throughout its history, the Museum has 
pursued joint missions of science and education. The Museum's power 
to interpret wide-ranging scientific discoveries and convey them 
imaginatively has inspired generations of visitors to its grand 
exhibition halls and educated its visitors about the natural world and 
the vitality of human culture. 

The Museum collections include over 32 million objects and 
specimens relating to anthropology, zoology, paleontology and the 
physical sciences. From the collected stories, clothing and material 
culture of the peoples of the Pacific Northwest Coast to the dinosaur 
eggs unearthed in Mongolia, from meteorites brought back from 
Greenland on a wooden sailing ship by Robert Peary to one of the 
most comprehensive sets of fossil horses ever assembled, the breadth 
and variety of the collections is astonishing. 

Illustrating the work of the Museum scientists and staff are over 
1.5 million black and white negatives, color transparencies, lantern 
slides and photographic prints held in the AMNH Research Library. 
The majority of these photographs were taken by Museum explorers 
and scientists who documented their field work. At the same time, 
they photographed the local environment and the people who lived 
there. The physical growth of the Museum was recorded by staff 
photographers who also photographed the work of the exhibition and 
education departments. These images are requested by researchers, 
students, educators, and professionals from around the world for 
academic and general publication, educational broadcast and 
distribution, artists' reference and personal use. 

The dissemination of this extraordinary Photographic Collection 
can be traced directly to Museum founder Albert Bickmore, one of the 
earliest and most enthusiastic advocates of visual education. 
Bickmore's lantern slide lectures were so successful that a new and 
larger theater was built in 1900 to accommodate the lines of teachers 



244 Picturing the Museum 



awaiting admission. To expand the Museum's educational mission 
beyond its walls, Bickmore created a lantern slide lending library of 
over 140,000 slides. The slide library formed the basis of the Natural 
Science Study Collections that were delivered to schools throughout 
New York State, (see Figure AMNH-i) 




Figure AMNH-1: Image 37244, Teachers selecting lantern slides; Image 313944, 
School Delivery, American Museum of Natural History Photographic Collection. 



Picturing the Museum 245 



In addition to the slides, the Museum delivered specimens and, 
later, model dioramas accompanied by lectures prepared by the 
Museum's educational and scientific staff. Bickmore's initiative 
foreshadowed the digital distribution of images by over a century. It is 
now possible to create worldwide access to the Collection and the 
public has come to expect to be able to find images online. Organizing, 
describing and digitizing a photographic collection of this size, 
however, is an enormous undertaking. 

Picturing the Museum 

To develop internal support within the institution for digitizing the 
Photographic Collection, the Library staff selected 50 appealing 
images related to the Museum's history and designed a small print- 
on-demand photo book to distribute at the January 2007 meeting of 
the AMNH Trustees' Library Committee (see Figure AMNH-2). 




Figure AMNH-2. Front and back covers of Picturing the Museum (© 2007), American 
Museum of Natural History, produced using Blurb.com (www.blurb.com). 

At the same time that the book was being developed, the Library 
used a similar theme, Natural Science Education and Exhibition at 
the American Museum of Natural History, as a digitization project 



246 Picturing the Museum 



proposal for METRO'S Regional Bibliographic Databases and 
Interlibrary Resources Sharing Program. This choice was made 
because the parameters of the grant proposal indicated that the 
subject matter should relate to the New York City Metropolitan Area. 
While most of the AMNH photographs were taken in distant places on 
every continent, the images of the Museum and its work were made in 
New York City and the surrounding area. 

The METRO funded project, Picturing the Museum: Education 
and Exhibition at the American Museum of Natural History (URL: 
http://images.library.amnh.org/photos/index.html) was created to 
showcase nearly 1000 images from the Photographic Collection, and 
to act as a prototype for developing a larger more comprehensive 
image database that would grow over time. The photographs in 
Picturing the Museum illustrate how the Museum staff presented 
natural science to the New York City public and the reactions of that 
public, particularly the school children. The pictures have been used 
by scholars studying education, museology, visual representation and 
local history. An NYU professor e-mailed the Library describing how 
the web site generated a new research topic, an article entitled 
"Exhibiting the Exhibit-Makers" about the history of the photographs 
of exhibition preparation at the AMNH. She is presently researching 
the Museum Archives to find more about the long-neglected Museum 
staff photographers in order to define their role historically within the 
institution. The pictures on the web site, documenting dinosaurs to 
fashion history, also tell a story about the people of New York City as 
the twentieth century emerged. 

Building upon Albert Bickmore's prescient vision of distributing 
images as an educational resource, the long term goal of the project is 
to make more images in the Collection readily available to researchers, 
scholars, students and the general public. Online access to the 
Photographic Collection will expand Bickmore's initiative 
exponentially, vastly increasing the numbers of those who will use the 
images for research, publication, broadcast and personal use. Students 
will be able to use these images in their class presentations and more 
individuals will be able to find pictures of their family from previous 
generations, like the Ainu woman from Japan, whose grandfather's 



Picturing the Museum 247 



photograph from the 1904 St. Louis Exposition is in the AMNH 
Photographic Collection. Online access will also help to preserve the 
original materials. Before scanning was institutionalized at the 
Museum, negatives would be repeatedly pulled from the collection for 
printing in the Museum's photographic studio. Now they are scanned 
once and left undisturbed, preserved in a climate controlled 
environment. 

Perhaps the most exciting potential of digital access to the AMNH 
Photographic Collection is that it offers the possibility of restoring 
provenance to many of the images in the Collection. With negatives 
filed numerically upon accession and prints arranged by subject to 
facilitate teaching, many collections were disassociated from their 
history. Study of historical photographic evidence is infinitely more 
valuable when examining a series of photographs instead of a single 
photograph out of context. 

Sources for the Image Metadata 

The basic records for the Picturing the Museum database were 
derived from 23 typed logbooks that had been created when the 
negatives were originally added to the Museum's collections. The 
logbook data reflects but does not consistently duplicate the data on 
the negative envelopes or on the photo print file cards that have been 
used to provide access to the images (see Figure AMNH-3). 

As part of a previous Mellon funded digital library project, the 
logbooks were scanned and then triple-keyed (a process by which the 
data was entered by hand three separate times and compared for 
discrepancies), an outsourced task. The result was 186,000 raw legacy 
records. Despite over a century of errors, anomalies, and 
abbreviations, using the legacy data as a source and editing and 
correcting that data is much faster than new cataloging. The verbatim 
legacy data itself is maintained in a separate field in the record. While 
all the data for Picturing the Museum came from the logbooks, it was 
important to plan the database design and write cataloging rules to 
consistently accommodate the larger Photographic Collection in all its 
variety. Fields not necessary in Picturing the Museum and later added 



248 Picturing the Museum 



to capture additional data for other collections included English 
translation, common name, scientific name, and cultural context. 
Images from rare books, expedition photographs, and images of 
artifacts would have this information associated with them. 




Figure AMNH-3. Data sources: logbook, back and front of photo file card, and 
negative sleeve, American Museum of Natural History Photographic Collection. 



Staffing and Workflow 

The METRO grant enabled the Museum to employ a part-time 
cataloger and a scanning technician for the project. Already on staff, 
the AMNH digital services librarian created a prototype web site for 
the image repository using Ruby on Rails and MySQL and the digital 
photographer created the web site's graphic design and styling. The 
Museum archivist managed the project and staff - consisting of 
volunteers, project employees, interns, and regular staff working on 
the project. The acting director provided expert direction and 
oversight. A library school intern with experience working for a stock 
photo agency did the image selection by reviewing the photo file cards 
and contributed greatly to the organization of the four themes chosen 
for the site: Dioramas, Education, Exhibition, and Exhibition 
Preparation. Workflow was planned so that scanning began 
approximately four weeks before cataloging to maintain the 



Picturing the Museum 249 



arrangement of the materials as organized for the project. The team 
met regularly to exchange materials and discuss the ongoing 
development of the database. 

Selection and Staging of Materials 

The intern choosing the images reviewed the photographs on the 
photo file cards and chose the most compelling. The negatives were 
scanned and uploaded to match the logbook data which had been 
imported into the database. Lastly, the cataloger edited the logbook 
data using the additional information found on the photo file cards 
and negative envelopes. 

Retrieving the negatives and staging the physical materials for 
Picturing the Museum required detailed planning and careful record- 
keeping. Several individuals worked with the materials to select, 
retrieve, scan, catalog, and re-file the images at different times. 
Project staff members in teams of two retrieved the negatives from the 
film storage room and grouped them by format to expedite scanning. 
Digitizing, describing, organizing, storing, and building a database for 
approximately 1000 images and their metadata was a large and 
complicated undertaking with a sizable team and many detailed 
procedures. The experience gained will be crucial for the ongoing 
digitization of the larger Photographic Collection. 

Equipment and Technical Specifications 

The funding received from the METRO grant allowed for the purchase 
of new equipment and software. For the project, the Library 
purchased the Epson Perfection V750-M Pro flatbed scanner, a MAC 
Pro work station, Eizo CG211: ColorEdge Color Calibration LCD 
Monitor with hood, and the Eye One Calibration hardware. Later in 
the project, due to an issue with moire patterns appearing on 26 
images, an anti-newton ring glass was purchased to rescan the 
problem images. Protocols on how to maintain a consistent color- 
managed environment, in order to ensure the tonal fidelity of the 
scanning and post-processing work were developed. An initial 
monitor profile was created with the Color Navigator software and the 



250 Picturing the Museum 



Eye-One Calibration device and the monitor was measured against the 
profile monthly to ensure that the calibration had not drifted. 

Most of the negatives scanned for this project were glass plate. 
The black and white negatives were captured in 16-bit grayscale TIFF 
files at an archival size equivalent to those recommended in the 
National Archives and Records Administration's manual, Technical 
Guidelines for Digitizing Archival Materials for Electronic Access: 
Creation of Production Master Files - Raster Images. All archival 
images were scanned to leave film borders and glass plate edges 
intact. The scans were corrected conservatively to make available as 
much of the information from the original negatives as possible. 
Manual dust and scratch removal was carried out when these 
interfered with image viewing. After the tone and texture correction, 
all images were converted to 8-bit grayscale and saved as TIFF files 
with borders cropped for web presentation on the Museum's SAN. The 
JPEG web derivatives are automatically uploaded and sent into 
storage alongside the TIFFs. 

The post-processing work procedures, based on the National 
Archives and Records Administration's best practices, are 
documented on the Picturing the Museum web site. Additionally, the 
technical information section of the project web site includes a 
detailed guide created by the Library's digital photographer, 
http://images.library.amnh.org/photos/imageworkflow.html. 

Conditions of Use 

The American Museum of Natural History endorses a policy that 
makes images in the Photographic Collection freely available to the 
research community to the fullest extent and as soon as possible. At 
the same time, it is recognized that access to some images must be 
restricted to protect the privacy of individuals and to respect various 
cultural traditions. An individual's image will be removed from the 
site upon request from that individual. Non-commercial users may 
download the files for their own use provided that they cite the 
Museum as the source of the image. No one may alter or modify the 
content without written permission from the Museum. 



Picturing the Museum 251 



Terms for commercial use of the images are detailed on the 
Picturing the Museum web site. These terms were written and 
subsequently approved by the Museum's General Counsel and 
composed so that they would refer to the Museum's Photographic 
Collection as a whole not just to the Picturing the Museum site. 

Development of the Database System 

The database for Picturing the Museum was designed to include 
search and browse features and an efficient cataloging interface, and 
to enable sharable metadata for interoperability. Metadata values and 
collection tables for the Picturing the Museum database were loosely 
modeled on the DSpace version 1.41 database schema. This design 
offered an opportunity to create multiple collections with distinct 
metadata schemas. Because the Library had different fields for 
different collections, instead of creating separate tables, metadata 
values were appended to a single table with a unique key to identify 
the field names for each collection. This design offered streamlined 
storage and a simple approach to dealing with repeating fields. The 
Library used MySQL to manage the images and metadata because it is 
a widely used database environment in the open source community, 
and was, at the time, the primary production database for Ruby on 
Rails web sites. 

Preliminary data clean-up of the logbooks was conducted using 
SOLR which is based on Lucene, a Java implementation of an 
enterprise search and retrieval library (Welcome to Solr, 2010). The 
Ruby-based SOLR framework was meant to be used as an advanced 
search mechanism across all collections, but the time to implement 
this feature became an issue. The web, however, is just one aspect of 
the Picturing the Museum image repository; Ruby is the language 
used to develop the computer workflow and image processing. 

Since the Library had extensive existing descriptive metadata 
comprising the bulk of a cataloged record, a scripted workflow was 
created with the Ruby language to ingest this metadata into 
corresponding collections, with a field indicating each specific logbook 
for each image. Metadata for these collections was ingested into the 



252 Picturing the Museum 



database prior to images being uploaded and a separate workflow was 
created in Ruby to upload images into the repository. During this 
upload, images were associated with a collection and a TIFF image 
and corresponding service master copy, a cropped original, and three 
derivative thumbnails were processed. Embedded rights and technical 
metadata using Harvard's open source Jhove were then extracted. 
Finally, an external library was created in Ruby to manage the 
retrieval and display of records in the web presentation environment. 

Cataloging Procedures and Metadata 
Documentation 

Editing and producing the photo book helped to refine the scope of 
the digital project and revealed some descriptive cataloging issues that 
would have to be addressed in the creation of an image database. For 
example, when producing the book we found that the original captions 
from the logbooks often did not translate into a good descriptive title. 
Images such as 22847, Mr. Albert with loan collections for school use, 
had no logbook caption at all. Images such as 274856, now reads Two 
boys collecting turtles, Trailside Museum, but was originally, Two 
boys collectors bring ing in turtles, including the grammatical errors 
and typos. Titles were composed to be as descriptive as possible using 
key words to increase retrieval. 

The descriptive data fields were formed according to DACS (SAA: 
Describing Archives: A Content Standard, n.d.) and mapped to 
Dublin Core (see DCMI Metadata Terms, 2008), unqualified, for 
future interoperability through OAI web harvesting. The descriptive 
fields for Picturing the Museum include the following: 

• Image Number 

• Medium (Source Format) (drop down based on the Thesaurus 
of Graphic Materials) 

• Size 

• Original Caption 

• Title 

• Original Photographer (drop down) 

• Copy Photographer (drop down) 



Picturing the Museum 253 



• Date 

• Geographic location (based on the Thesaurus of Geographic 
Names) 

• Person 

• Institution 

• Permanent Hall 

• Expedition 

• Temporary Exhibition 

Roughly based on the VRA Core Element Description (see 
Welcome to VRA Core 4.0, n.d.), syntax rules for the title field were 
crafted to be brief and consistent. Writing a new title but retaining the 
original caption helped maintain integrity to the original data and the 
history of the Collection, while allowing for a broader search. For 
example, an original caption of "Corean hats" would be edited to 
"Korean hats" in the title. A search by either spelling would yield the 
record. Also, original captions reflecting social and political realities of 
the time of the photography would be retained while a more accurate 
or culturally sensitive title could be crafted. For example, the larger 
database includes many images of artifacts from the Congo; Congo is 
retained in the source data, while the geographic location field now 
reads Congo (Democratic Republic), as it appears in The Getty 
Thesaurus of Geographic Names (i.e., TGN, see Getty Thesaurus of 
Geographic Names Online, n.d.,) which was used to create the 
standard names of places in this field. Similarly, cultural names such 
as "Indian" would be changed to "Native American", "natives" to 
"people" or "people of ...." Researchers can therefore find information 
by both historical and present-day terms retaining the original 
documentation. 

Considering the size of the AMNH Photographic Collection as a 
whole, the records were designed for easy and fast data entry that 
would comply with documented standards for interoperability. 

The geographic location field wasn't as much of a challenge 
during Picturing the Museum, but when the database was expanded to 
include field photographs and images of artifacts, the metadata 
became more complex. For example, a 19th century figure of Burmese 



254 Picturing the Museum 



jadeite carved in China required a repeated field for geographic data 
(as well as an updated geographic name of Myanmar for Burma). 

Because there are so many more place names in the AMNH 
Photographic Collection than appear in the TGN, the cataloger uses 
the broadest term in the string in the Geographic Location field and 
references the specific site or place name in the title or note field. 
However, the AMNH Library plans to contribute these place names, 
such as Hopetown, British Columbia, to the TGN in the future as the 
Photographic Collection contains images from many smaller villages 
and localities worldwide. 

Information that isn't suitable for any other field and that isn't 
found in the TGN or the Library of Congress Authorities but is 
necessary for describing the image is placed in the title or note field so 
that all data is captured and will appear in search results. 

Working on the book also highlighted the need for local 
authorities for photographer's names, permanent halls, temporary 
exhibitions, and Museum expeditions. The first developed was the 
photographer list. Using SOLR, the programmer produced a unified 
list of the many name variations in the 23 logbooks. Names were 
validated in the catalog and files of the AMNH Research Library as 
well as through the Library of Congress Name Authority Headings 
then inserted into a drop-down box on the cataloger's interface. 

For over a century of photographs, there were over 262 
photographer names, including celebrated individuals such as Roy 
Chapman Andrews and Edward S. Curtis. For many images, no 
photographer was listed at all and "Unknown" was entered. Authority 
lists are also required for temporary exhibits, permanent halls, and 
expeditions. Not only are these terms unique to the American 
Museum of Natural History, but they have changed over the years 
prompting discussions about whether to use current Museum terms 
or historic ones to reflect the Museum the way it is today, or the way it 
was when the images were photographed. 

Choosing metadata fields and creating a data dictionary that 
reflects the content in all collections as well as in Picturing the 
Museum requires planning and flexibility. 



Picturing the Museum 255 



Database Growth 

Because the image database continues to grow, continuity in staffing 
and procedures has proven valuable. A guide to cataloging procedures 
is available on the web site (see PTM Cataloging Procedures, n.d.) 
Expanded examples and data fields for other Collections are 
documented internally. The cataloger recorded issues throughout this 
entire process allowing the programmer to make improvements to the 
interface to accommodate the metadata needs of different collections. 
Documentation also helped to prepare for the training of student 
interns and volunteers to catalog in the database. 

The cataloging interface for the project evolved over time and 
continues to change as additional image collections are digitized and 
cataloged. For example, a 1928 hand-colored lantern slide from a 
Museum expedition to Mongolia will require different metadata than 
images selected for Picturing the Museum due to medium, the need 
for additional local subjects, and a field for Museum expedition, and 
there are many such items in the Collection. 

Fields added to the database after Picturing the Museum include 
the following: 

• Artist (drop down) 

• Publication 

• Catalog/Specimen Number 

• Object/ Specimen Location 

• Department/Discipline (drop down) 

• Cultural Context 

• Common Name 

• Scientific Name 

• Local Subject (drop down) 

Drop down lists are used for ease and consistency in fields like 
original and copy photographer, artist, department/discipline and 
local subject. This last term allows for a somewhat standardized 
vocabulary that can be searched for often used subject terms. For 
example without this data field searching for "birds," would not yield 
an image titled "Cardinal." Images of dinosaurs often don't use that 



256 Picturing the Museum 



word in the title e.g. "skull of Tyrannosaurus Rex," would not be found 
by searching "Dinosaur." This approach will yield a large if not 
comprehensive data set for the general researcher. More precision will 
result from geographical headings and fields with authority controlled 
vocabularies. Drop down lists will be added when the authority work 
is completed for Permanent halls, Temporary exhibits and Museum 
expeditions. 

Lessons Learned and the Future of the Database 

The type of images in Picturing the Museum — historical documentary 
photographs generally taken in the Museum and the New York City 
area — are a subset of the AMNH Photographic Collection. Other 
kinds of pictures coexist within that larger collection: field 
photographs taken on Museum sponsored expeditions, images of 
specimens and artifacts, and portraits of individuals. Authority work 
is a current priority, particularly for names of Museum halls and 
Museum expeditions. Lists are nearly completed for permanent halls 
and temporary exhibits but creating authorities for Museum 
expeditions is a larger and more complicated task and will require 
some time. 

Picturing the Museum is the only portion of the Image Database 
available to the public at this time but the entire database is searched 
frequently by the Library staff to fulfill research requests from both 
inside and outside of the Museum. This AMNH Image Database 
combines the logbook data, another smaller color database as well as 
sets of data specific to collections that were not integrated into the 
larger numbering scheme. The staff continues to make documentation 
of procedures and data standards a priority, especially as the database 
grows to include the metadata fields for other types of collections, e.g. 
specimen or artifact numbers, artists' names and the names of culture 
groups. Work is proceeding with the help of volunteers. The part time 
image cataloger for the METRO project, now a permanent staff 
member, is developing a training program for volunteers and interns 
to edit the catalog records based on a similar plan using volunteers to 



Picturing the Museum 257 



scan and process images. The AMNH Library has close to 60,000 
images scanned and over 200,000 largely unedited records. 

Researchers request photographs on a routine basis and even 
though most of the records are unedited the database has become an 
indispensible time saving tool to search for an image and then 
determine whether or not it has been scanned. Providing individual 
images for use in books, articles, classroom projects and videos is a 
mission related service of the Library. However, it is also important to 
recognize that this is a photographic archive and for scholarly research 
projects, it is essential to view the images together in their historical 
context. Despite a century long tradition of item level description, one 
of the major opportunities provided by the Image Database will be to 
recollect the collections of images that have been dispersed in the 
Museum collections over time and place them in their historical 
context. 

The Library has begun a prototype of a web template to be used as 
an interface to contextualize collection specific information. By 
identifying and grouping the images, e.g. by the name in the 
expedition field, it will be possible to retrieve a preselected group of 
images from that expedition and link to a select dataset as the 
container list on the web site. The other text, e.g. site overview (scope 
and content); biographical notes, historical notes, bibliographies, 
other resources map can be easily encoded into an EAD (see Encoded 
Archival Description Version 2002 Official Site, 2009) finding aid. As 
more of these special photographic collection web sites are developed, 
the Library will create a separate interface on its home page for direct 
access to these collections. As part of that effort, it will also integrate 
the means for viewers to comment on the images, members of 
originating cultural communities, scholars, students and others will be 
able to correct and add to the information currently associated with 
the images, expanding the knowledge surrounding their history. 
Mapping to EAD and at the same time making the collection level 
records for the web sites harvestable by OAI, will also increase the 
possibility of retrieval of these collections. 



258 Picturing the Museum 



In a world of technological change and seemingly limitless 
possibilities for communication, the AMNH Library is expanding on 
Bickmore's vision for sharing visual resources and knowledge for the 
people for science, and for education, now and in the future. 

References 

PTM Cataloging Procedure, (n.d.). Retrieved May 1, 2010 from 

http://images.library.amnh.org/photos/PTM_CatalogingProcedur 
es.pdf 

DCMI Metadata Terms. (2008). Retrieved May 1, 2010 from 
http : / / dublincore. org/ documents / dcmi -terms / 

Encoded Archival Description Version 2002 official site. (2009). 
Retrieved May 1, 2010 from http://www.loc.gov/ead/ 

Getty Thesaurus of Geographic Names Online, (n.d.). Retrieved May 
1, 2010 from http://www.getty.edu/research/ 
conducting_research/vocabularies/tgn/ 

SAA: Describing Archives: A Content Standard (DACS) (n.d.). 
Retrieved on May 1, 2010 from http://www.archivists.org/ 
governance/ standards/dacs.asp 

Welcome to Solr. (2010). Retrieved March 30, 2010, from 
http://lucene.apache.org/solr/ 

Welcome to VRA Core 4.0. (n.d.). 

http : / / www. vraweb . org/pr oj ects/ vracor e4 / 



Part III - The Digital 
Campus: Digitization in 
Universities and Their 
Libraries 



259 



Developing an Institutional Repository at 
Southern New Hampshire University: Year 
One 



Alice Piatt (Southern New Hampshire University) 



Abstract 

In 2008, Southern New Hampshire University was awarded a three- 
year, $500,000 national leadership grant from the Institute of 
Museum and Library Services to create a digital repository using 
DSpace open source software. Events from the first year of the 
repository's development are presented and discussed. Key elements 
addressed include the challenges involved with customizing the 
DSpace infrastructure, creating standards for access and master files, 
implementing metadata standards, and developing digital 
preservation policies. The value of cross-departmental participation is 
shown, and the importance of planning for digital preservation is 
presented. 

Keywords: Best practices, Digital library, ETD, Electronic, 
Interfaces, Institutional repository, Open source, Scans. 



Introduction 

In 2008, Southern New Hampshire University was awarded a three- 
year, $500,000 national leadership grant from the Institute of 
Museum and Library Services (IMLS) to create a digital repository 
using DSpace open source software. The inspiration for the project 
was a collection of student theses and dissertations from the School of 



261 



262 Developing an Institutional Repository 



Community Economic Development (SCED). SCED is a unique 
program with participation from all over the world, particularly the 
United States and Tanzania, and also countries such as Uganda, Peru 
and the Philippines. Like many thesis collections, the projects were 
printed using consumer-grade equipment, and only one copy was 
bound and saved for the library. The international nature of the 
projects, in addition to the danger of losing them to deterioration, 
made them an attractive collection for beginning a digital repository. 
Faculty papers from the International Business program were also 
included in the grant project, to make papers once only accessible 
from a professor's office available to the world. 

Many institutions lack the financial and human resources to build 
a successful digitization program. The gap between resources available 
versus resources required can often be bridged by a grant; a search for 
IMLS grants from 2004-2008 using the keyword "digitization" shows 
that at least 57 IMLS grants were provided to libraries and museums 
for digitization projects (IMLS, 2009). Like most institutions, the 
repository at SNHU's Shapiro Library could not have come to fruition 
without grant assistance. 

The following pages share the Shapiro Library's experiences 
during the first year of repository development. 

The People Involved 

Digitization programs need a strong level of organization and 
administrative support to succeed. Programs that only live within the 
walls of the library without buy-in from administration and other 
departments are at risk of failure for lack of support. The Shapiro 
Library's digital repository is managed by a Digital Initiatives 
Librarian, who receives support from the Digital Content Specialist, 
two graduate assistants, and two cross-departmental committees: the 
Implementation Committee and the Policy Team. 

The Digital Initiatives Librarian is responsible for managing the 
repository, including creation of metadata standards, scanning 
workflows, policy development, and quality control. The Digital 
Content Specialist creates descriptive keywords, and writes abstracts 



Developing an Institutional Repository 263 



for the theses. Two graduate assistants were hired to execute the 
scanning, optical character recognition (OCR), and access file 
creation. 

The Implementation Committee was initially organized to 
prepare the grant application, and after the grant was received, 
organized the necessary infrastructure. Represented on the committee 
are the Library Dean, the Electronic Resources Librarian, the 
University Webmaster, the Dean of the School of Community 
Economic Development, and both the head of the IT department and 
the IT programmer committed to the project. The committee hired the 
Digital Initiatives Librarian and a Digital Content Specialist, who both 
subsequently joined the committee. The Implementation Committee 
continues to meet on an as-needed basis to monitor the repository's 
development. 

While some members of the Policy Team are consistent with the 
Implementation Committee, the focus for this group is to determine 
policies for the repository and discuss other questions that might 
arise, whether they are related to file format, collection development, 
or metadata. Because of the nature of the team, there are more 
librarians represented: the Electronic Resources Librarian, Technical 
Services Librarian, and the Access Services Librarian are all part of the 
team, as well as the Digital Initiatives Librarian, Digital Content 
Specialist, and the Library Dean. Also on the team are the IT 
programmer and the Associate Dean of the Faculty. The associate 
dean's participation is effective in keeping the university 
administration informed on the progress and policies of the 
repository. The Policy Team initially met every two weeks, and 
continues to meet at least once a month. 

Developing the Technical Infrastructure 

After identifying the initial collections for the repository, the 
Implementation Committee selected the digital repository software 
and the hardware on which it would reside. Oya Y. Rieger (2007) 
explains that when selecting software, a number of factors should be 
considered, including matching your institution's needs to the 



264 Developing an Institutional Repository 



software's features, considering what resources will be required to 
install and maintain the software, and assessing the overall usability 
for both staff and end-users. Often the question might arise: to open 
source, or not to open source? While using open source software is the 
current trend, institutions should look closely at their resources to 
determine if they can support the technological and human resources 
required to work with open source software packages. 

The grant awarded to SNHU included funds to hire the Digital 
Initiatives Librarian and the Digital Content Specialist. It also 
financially supported time spent working on the grant by other 
positions already in place, including IT. Assessment of these resources 
determined that enough support was available to consider open 
source software. DSpace stood out as the most widely-used open 
source institutional repository software package available for 
academic library use, with an active user community and a wide array 
of resources available (DuraSpace, 2009, Resources). 

The differences between implementing open source versus 
proprietary software quickly became apparent. While DSpace is 
advertised to be useful "out of the box," this is not a realistic assertion 
(DuraSpace, 2009, About DSpace, para. 1). A certain level of 
programming skill and time is required in order to customize the 
software. In DSpace, the level of programming needed to make 
customizations beyond changing the color scheme of the website can 
be daunting for someone without experience in both programming 
and website design. The Digital Initiatives Librarian's web design 
skills and the IT programmer's skills were both needed to make most 
of the necessary customizations to the user interface. 

Community support - While the DSpace community is very 
active, with a well-populated wiki and listservs for general and 
technical questions, it is also a complex community. Users vary by 
what platform they work on (Linux vs. Windows) and what version of 
DSpace they use. During the time of SNHU's installation, most of the 
user community was working with either DSpace 1.4 or 1.5. To further 
complicate things, some users of 1.5 were using what is known as the 
JSP user interface, while others used the XML user interface - each 



Developing an Institutional Repository 265 



involving different programming methods for customization. 
Therefore, not all questions and answers posted by the community are 
relevant to one's needs. One example encountered was a DSpace wiki 
entry explaining how to change the DSpace code to enable linking 
authors in a simple item record. When the code did not function 
properly, the question was posted to the DSpace tech listserv. Another 
community member explained that the encoding described in the wiki 
had changed in version 1.5 (Piatt, 2009). Additionally, answers to 
questions regarding installation varied widely depending on if the user 
was on Linux or Windows. While DSpace does have a large user 
community, that community requires some careful navigation. 

Professional development - In early June 2009, the NITLE 
consortium presented a timely DSpace workshop (NITLE, 2009). The 
variety of sessions provided a strong background to DSpace's 
capabilities. One session in particular, "Developing Interfaces and 
Interactivity for DSpace with Manakin Workshop" by Eric Luhrs of 
Lafayette College, was extremely helpful, providing tools and the 
know-how necessary to make customizations to the XML user 
interface (Luhrs, 2009). Without the benefit of this interactive 
instruction, the learning curve involved would have been much more 
difficult to transcend. 

The experience at the DSpace workshop points to the importance 
of this type of professional development in the rapidly-changing 
digital library environment. Conferences such as the Open 
Repositories Conference and the Joint Conference on Digital Libraries 
have both included specific DSpace sessions and workshops in the 
past. A simple search of the web reveals user groups and workshops 
available for other digital library platforms, including proprietary 
software such as OCLC's CONTENTdm. Providing funding for 
librarians and IT staff to attend these types of educational events 
should be a priority for any institution embarking on a digitization 
project. 



266 Developing an Institutional Repository 



From Paper to Electronic 

Creating an electronic record for access involves metadata authoring, 
scanning, and access file creation. 

Metadata - Metadata standards should be determined before 
the first item is ever added to the repository. Because qualified Dublin 
Core is installed with DSpace by default, and because Dublin Core is 
the leading schema for describing digital resources, it was selected for 
the schema. Determining which elements to make available in the 
DSpace submission form was more challenging. Not every element 
should be used to describe a digital object - not all are appropriate for 
all collections. Besides, the time-consuming nature of metadata entry 
requires that standards be chosen with efficiency in mind. Michael 
Boock and Sue Kunda (2009) explain how creating a metadata record 
for both the DSpace repository and MARC catalog can take up to an 
hour per record, even when students create the majority of the 
descriptive metadata (p. 300-302). While it is important to consider 
descriptive, administrative, structural, and preservation metadata, 
these elements must be chosen carefully to achieve thorough, but cost- 
effective item description. 

The CDP Metadata Working Group's "Dublin Core Metadata Best 
Practices", the DCMI Usage Board's "DCMI Metadata Terms", the 
Scholarly Works Application Profile as described by Julie Allinson 
(2008), and the Networked Digital Library of Theses and 
Dissertations' metadata standard (Atkins, Fox, France, & Suleman, 
2008) were all examined. From these best practices, 32 qualified 
Dublin Core elements were selected, with the intention that any item 
added to the repository could be appropriately described using some 
or all of these elements. Approximately 20 of these are used to 
describe the SCED thesis projects in particular. 

Scanning - During the development of the DSpace 
infrastructure, the scanning workflow was also launched. The initial 
collection of student theses and dissertations from SCED proved to be 
challenging to scan. Part of the purpose of the SCED thesis project is 
to document work completed by the student in the field, outside of the 
classroom. To that end, most of the theses, collected from 1984- 



Developing an Institutional Repository 267 



present, include large appendices of documentation including letters, 
financial statements, marketing materials, photographs, architectural 
plans, and even a wall calendar used as a fund-raiser. Additionally, 
students were given the opportunity to be creative in their 
presentation, often using color, graphs, and decorative fonts. 

Sample theses were selected and scanned by the graduate 
assistants to test how the scanner and OCR software would handle the 
diverse materials. These initial scans immediately raised questions. 
There were not yet policies in place for how much information should 
be captured in the scans, causing uncertainty when incidental color 
was encountered, such as flyers printed on colored paper. 
Additionally, there was confusion surrounding the fact that the 
digitization process includes preserving master files, saved in 
traditional TIFF format, in addition to the PDF files created for access. 
The IT staff was not prepared to store and preserve this large 
collection of master files, and panic arose about their massive size - 
the files, scanned at 600 dpi in grayscale or color, were 30 to 80 
megabytes each. This "megabyte shock" is not unusual, particularly at 
small institutions; Stacy Nowicki (2008) also noted problems with 
large TIFF files at Michigan's Kalamazoo College. 

After much discussion, the Policy Team agreed to scan the papers 
for their intellectual content only. Best practices from the California 
Digital Library (2008) and the CDP Digital Imaging Best Practices 
Working Group (2008) were consulted to determine digitization 
standards: a 600 dpi setting for black and white pages, and 500 dpi 
for grayscale or color pages. This 500 dpi setting resulted in a 
minimum of 4000 pixels on the largest side of the scan, in accordance 
with these recommendations (California Digital Library, 3.6.1; CDP, p. 
8). It should be noted that if the pages were significantly a different 
size, the dpi setting would be adjusted to meet this parameter. Master 
files are saved in TIFF format (California Digital Library, 2008, 3.2). 
Grayscale and color are only used when necessary to preserve the 
intellectual content of the document, leaving most of the pages to be 
scanned as black and white. As a result, the master files are much 
smaller; the black and white scans are approximately 4 megabytes 
each. 



268 Developing an Institutional Repository 

The solution to the color question requires a certain amount of 
human judgment, but is viable because the Digital Initiatives 
Librarian and the graduate assistants conducting the scanning are 
located in close proximity to one another, facilitating an environment 
for quick decisions. David Lowe and Michael Bennett (2009) state 
that the Internet Archive chose to scan all their documents in color, 
eliminating the need for human judgment (p. 210). 

Access File Creation - After creating master TIFF files, it is 
necessary to convert them for public access. The Portable Document 
Format (PDF) format, processed so that full-text searching is possible, 
is ubiquitous among subscription and open access academic 
databases. It was the obvious solution for our collection. 

To enable full-text searching, the TIFs were processed using 
optical character recognition (OCR) software. ABBYY FineReader 9 
Professional was selected, based on a review in PC Magazine 
(Mendelson, 2008). This feature-rich software enables OCR 
recognition and error-checking in multiple languages, and performs 
well with most text, including text printed with a dot-matrix printer, 
and text formatted in blocks, such as in newsletters and flyers. 

From FineReader, the graduate assistants are able to save the 
PDF with an option called "text under image," saving the corrected 
OCR text in an invisible, searchable layer under the scanned page 
image. In order to keep the size of the file reasonable, the PDF images 
are saved at 300 dpi; to enhance accessibility, the option for creating a 
tagged PDF is selected (Johnson, 2004). 

After the PDF is created, it is opened in Adobe Acrobat, and 
additional metadata is added to the file's properties, including title, 
author, and copyright status. 

Information in the repository should be not just available, but 
accessible to all. This includes maintaining file sizes to enable faster 
load times, ensuring that even users with dial-up modems can 
download the files in a reasonable amount of time. According to a 
survey led by John Horrigan at the Pew Internet & American Life 
Project (2009), seven percent of Internet users in the United States 
are using dial-up services at home (p. 7). While seven percent sounds 



Developing an Institutional Repository 269 



small, it is equal to approximately 9 million households in the United 
States, out of the 129 million counted by the U.S. Census by July 1, 
2008. International user statistics vary widely, but it would be best to 
avoid frustrating any users with unnecessarily large file sizes, thus 
increasing the viability of the collection. 

Therefore, nearly all of the projects are split into two PDFs. 
Because the bulk of most of the thesis projects is the supporting 
documents in the appendix, the papers and their appendices are saved 
as separate PDF files. Of the first 88 student projects scanned, the 
average file size of the project paper by itself was 2.33 megabytes, with 
a median of 1.73. The appendices' average was 8.29 megabytes, with a 
median of 5.2. Both the main paper and the appendix PDFs are 
available from the same item record in DSpace. 

To improve the access files' longevity and accessibility, the PDFs 
are saved as PDF/A when possible. Roger Reeves and Hans Barfuss 
(2009) explain the International Standards Organization's (ISO) goal 
for PDF/A is that it "provides a mechanism for representing electronic 
documents in a manner that preserves their visual appearance over 
time, independent of the tools and systems used for creating, storing 
or rendering the files" (The Goal of PDF/A). One example of the 
advantage of PDF/A is ensuring that elements such fonts are 
embedded in the file, so they display properly even if the user does not 
have those particular fonts already loaded on his computer. The utility 
provided in Adobe Acrobat Professional 9 was used to save the files in 
PDF/A format. 

The Often-Missed Point: Digital Preservation 

While all of the decisions involved with customizing the DSpace user 
interface were being addressed, one major component of the 
digitization program was not addressed: the concept of digital 
preservation. While it was understood that digital preservation was an 
issue, it was uncertain how preservation would be accomplished. 

The ICPSR Digital Preservation Workshop at the University of 
Michigan was an excellent opportunity to learn more about digital 
preservation. This five-day, in-depth workshop made it clear that if an 



270 Developing an Institutional Repository 

institution presents digital documents online, there is an assumption 
that they will be preserved there forever - much like a book on the 
shelf is expected to be readable ten, fifty, or even hundreds of years 
after it is bound. However, digital files are fragile in their own way, 
and are susceptible to obsolescence, storage media problems, and 
other issues (Cornell, 2007, Tutorials, chapter 3: Obsolescence & 
Physical Threats). 

While digital preservation is a complex topic with many 
components and considerations, the primary concern was how to 
adequately care for the master TIFF files. Each image must be 
preserved in the event that the access PDF file becomes corrupted, or 
when PDF is superseded by a new access file format. The Digital 
Preservation Tutorial developed by Cornell makes it clear that 
institutions can not burn files to CDs, put one CD on the shelf, another 
in someone's garage, and believe they have preserved their files 
(2007). According to the tutorial, even CD standards have changed 
over the years, and early formats are now obsolete (Chapter 3: 
Obsolescence, Chamber of Horrors, Disk Media). It is also apparent 
that backing up files without including any descriptive information is 
still not adequate preservation; how many of us have opened a floppy 
disk and wondered, "What the heck is all this stuff?" Master files must 
also have their own metadata associated with them to describe what 
they are. But learning how to preserve these files, as well as 
adequately preserving the access files and their associated metadata in 
these early years of digital preservation, is a challenging process that 
has not been adequately addressed during the first year of repository 
development at the Shapiro Library. It is probable that many other 
institutions have also not addressed their own digital preservation 
questions, or even asked them. 

The problems with file obsolescence and data backup are just one 
small component of creating a digital preservation program. The 
guidelines presented by the ICPSR workshop are a helpful resource in 
determining how to ensure that digital preservation at the Shapiro 
Library is compliant with standards described by the Reference Model 
for an Open Archival Information System (OAIS), an industry 
standard. Much progress is anticipated for the second grant year. 



Developing an Institutional Repository 271 



Conclusion 

The myriad of details involved with creating a digital repository at 
Southern New Hampshire University were more complex than 
anticipated. Learning and implementing standards for metadata, 
master files and access files was time-consuming, but taking the time 
to establish standards in the beginning doubtless saved a great deal of 
trouble for the future. Even so, it will be necessary to keep up with 
developing industry standards, and it would not be surprising if 
further adjustments are needed down the road. A digital repository is 
much like a physical building; periodic maintenance, remodeling, and 
wear and tear should be anticipated and expected. 

The Shapiro Library's digitization program has strong 
administrative support, participation from several university 
departments, and strong financial resources. The repository will 
become a successful program for the university long after the grant 
period concludes. 

References 

Allinson, J. (2008). Describing scholarly works with Dublin Core: A 
functional approach. Library Trends 57(2), 221-243. 

Atkins, A, Fox, E., France, R. & Suleman, H. (2008). ETD-MS: an 
interoperability metadata standard for electronic theses and 
dissertations, ver. 1.00, rev. 2. Retrieved from http://www. 
ndltd.org/ standards/metadata/ etd-ms-vi.00-rev2.html 

Boock, M., & Kunda, S. (2009). Electronic thesis and dissertation 
metadata workflow at Oregon State University Libraries. 
Cataloging & Classification Quarterly, 47(3/4), 297-308. 
doi:io.io8o/oi639370902737323. 

California Digital Library. (2008). CDL guidelines for digital images. 
Retrieved from http://www.cdlib.org/inside/diglib/guidelines/ 
bpgimages/reqs.html 

CDP Digital Imaging Best Practices Working Group. (2008). BCR's 
CDP digital imaging best practices version 2.0. BCR. Retrieved 
from http://bcr.org/dps/cdp/best/digital-imaging-bp.pdf 



272 Developing an Institutional Repository 



CDP Metadata Working Group. (2006). Dublin Core metadata best 
practices: version 2.1.1. BCR. Retrieved from 
http://vvrww.bcr.org/dps/cdp/best/dublin-core-bp.pdf 

Cornell University Library. (2007). Digital preservation 

management: Implementing short-term strategies for long-term 
problems. Retrieved from http://www.icpsr.umich.edu/dpm/ 

DCMI Usage Board. (2008, January 14). DCMI Metadata terms. 
Retrieved from http:/ / dublincore.org/ documents/ dcmi-terms/ 

DuraSpace. (2009). DSpace. Retrieved from http://dspace.org 

Horrigan, J. (2009). Home broadband adoption 2009. Pew Internet 
& American Life Project. Retrieved from http://www.pewinternet. 
org/Reports/2009/ io-Home-Broadband-Adoption-2009.aspx 

Institute of Museum and Library Services. (2009). Search awarded 
grants. In Grant Search. Retrieved from http://imls.gov 

Johnson, D. (2004). What is tagged PDF? In Accessible PDF 
Learning Center. Retrieved from http://www.planetpdf.com/ 
enterprise/ article.asp?ContentID=6o67 

Lowe, D. B. & Bennett, M. J. (2009). A status report on JPEG 2000 
implementation for still images: The UConn survey. Archiving 
2009, 6, 209-212. 

Luhrs, E. (2009). NIS Camp: Developing interfaces and interactivity 
for DSpace with Manakin. Retrieved from 
http://nitlecamp.pbworks.eom/f/manakin-workshop-slides.pdf 

Mendelson, E. (2008). ABBYY FineReader Professional 9.0. PC 
Magazine. Retrieved from http://www.pcmag.com/article2/ 
o,28i7,230562i,oo.asp 

NITLE Information Services. (2009). Nitlecamp. Retrieved from 
http://nitlecamp.pbworks.com/ 

Nowicki, S. (2008). Using DSpace for institutional repositories. 
Retrieved from http://hdl.handle.net/10090/4522 

Piatt, A. (2009, Oct. 6). Two theme modification questions. Message 
and responses, archived at http://sourceforge.net/ 
mailarchive/forum.php?forum_name=dspace-tech 



Developing an Institutional Repository 273 



Reeves, R. and Barfuss, H. (2009). PDF/ A - A new standard for long- 
term archiving. PDF/A Competence Center. Retrieved from 
http://vvrww.pdfa.org/doku.php? id=pdfa:en:pdfa_whitepaper 

Rieger, O.Y. (2007). Select for success: Key principles in assessing 
repository models. D-Lib Magazine, 13(7/8). 
doi : 10. i045/july2007-rieger 

U.S. Census Bureau. (2009). Annual estimates of housing units for 
the United States and States: April 1, 2000 to July 1, 2008 (HU- 
EST2008-01). In Housing Units: State Housing Unit Estimates: 
2000 to 2008. Retrieved from http://www.census.gov/popest/ 
housing/HU-EST20o8.html 



Digitization of the Yale Daily News 
Historical Archive 



Kathleen Bauer, Ian Bogus, Karen Kupiec 
(Yale University Library) 

Jennifer Weintraub (UCLA Library) 
Abstract 

The Yale Daily News is Yale University's independent student run 
newspaper. Founded in 1878 it is the oldest continuously published 
daily newspaper at a United States university. From the initial print 
volumes until digital versions started in 2000, the entire run of the 
printed paper consists of 122 volumes and approximately 100,000 
pages. In 2007, Yale University Library was asked to create a pilot 
project to digitize and make available an initial set of ten years of the 
newspaper with a $50,000 start-up budget. In this article, we will 
discuss how the project began, and issues that developed during the 
process related to copyright, interface design, workflow, quality 
control, and fundraising. This project helped Yale University Library, 
a large, strongly hierarchical institution, to develop workflows that 
allow its staff to develop new skills and work across traditional 
departmental boundaries. Library staff that have traditionally 
performed tasks related to our print collections or for smaller digital 
projects have developed new skills and methods for workflow, 
metadata creation and quality control for a large-scale digital project. 

Keywords: Campus newspaper, Copyright, Newspaper digitization, 
METS, Quality control, Yale, Scanning. 



274 



Digitization of the Yale Daily News Historical Archive 275 

A newspaper digitization project is one that every library, public or 
academic, can undertake. It is often not hard to get the rights to a local 
or small paper and an academic or public library has a built-in 
audience for this type of project. Local researchers will love having it 
online and genealogists from further afield will bless you. And yet, 
newspaper digitization, while having recently come into its own, has 
been somewhat difficult for libraries. Newspapers are crucial to 
research, providing detailed local and international accounts of 
events; these incredibly important primary source materials are made 
of poor quality material that will last a relatively short period of time. 
Newspapers are hard to digitize because they are published daily with 
hundreds of issues a year, comprised of various individual sections, 
and then individual articles, oversized, delicate, and contain 
thousands of words and pictures that require careful quality 
assurance. In addition they have unusual layouts, and articles often 
are split across two or more nonconsecutive pages. There can be 
numerous contributing authors, syndicated cartoons, advertisements, 
supplements, and even the occasional joke issue. 

Fortunately, newspaper digitization is not new. Many 
organizations have taken on newspaper digitization and the major 
national and regional newspapers are now available for licensing by 
libraries. While many projects focus on digitization of newspapers 
from microfilm, there is also an increasing number of digitization 
projects that begin with the original paper. One important 
clearinghouse for information and best practices for newspaper 
digitization is the National Newspaper Digitization Program (The 
Library of Congress, 2009). This program, a joint effort between the 
Library of Congress and the National Endowment for the Humanities 
uses the power of grant dollars to enable proper newspaper 
digitization, research in newspaper digitization and access to the 
digitized papers through a central resource. 

The Yale Daily News, Yale University's student run newspaper, is 
132 years old and is the oldest continuously published daily 
newspaper at a United States university. In 2007, Yale University 
Library (YUL) was asked to create a pilot project to digitize the 
newspaper with an initial $50,000 start-up budget provided by the 



276 Digitization of the Yale Daily News Historical Archive 

Yale Daily News's parent foundation (the Oldest College Daily 
Foundation) and YUL. In this article, we will discuss how the project 
began, and issues that developed during the process related to 
copyright, interface design, workflow, quality control, and 
fundraising. 

This project helped YUL, a large, strongly hierarchical institution, 
to develop workflows that allow its staff to develop new skills and 
work across traditional departmental boundaries. Staff across the 
Library who have traditionally performed tasks related to our print 
collections or for smaller digital projects have developed new skills 
and methods for workflow, metadata creation and quality control for a 
large-scale digital project. 

The Yale Daily News (YDN) is staffed and produced by student 
volunteers. The paper is not owned by Yale University, and the 
student reporters and editors are advised by the independent Oldest 
College Daily Foundation (OCD). OCD is comprised of former YDN 
staffers and Yale graduates. In 2005 the OCD came to the YUL with an 
idea for a project to digitize the Yale Daily News archive and provide 
access on the Internet. The OCD realized the complexity of the 
proposal especially considering they did not own a complete run of the 
newspaper. They asked the YUL to partner with them as OCD owned 
the rights to the content while YUL had the expertise and the means to 
make it accessible. To start the pilot project, OCD and YUL 
contributed $25,000 each to finance a pilot project. YUL decided that 
for the pilot project we would not digitize anything for which there 
was an existing digital edition (the YDN has been available online 
since 2000). Thus, we still had to choose a small amount of material 
from 120 years of print issues, or a fraction of the 100,000 pages in 
the entire run, for our initial digitization pilot. 

This type of partnership, between an external group owning 
copyright and the campus library, can be useful for both parties. It is a 
good way for library staff to gain experience with a complex 
digitization project and digital collection building, it provides useful 
material for fundraising for technology projects, it enables the library 
to provide a useful resource to the campus community, and it enables 



Digitization of the Yale Daily News Historical Archive 277 



both the newspaper and the library to create an online product with 
research value freely for a product that may not have a large sales 
market. 

Several basic principals helped guide the development of the Yale 
Daily News Historical Archive. Open or commonly used standards for 
our digital files were important in the event content needed to be 
migrated to new interface software in the future. We wanted to 
digitize each newspaper in its entirety, thereby preserving the 
historical context provided by editorials, cartoons and advertisements. 
Therefore it was important that we capture the images on each page, 
not only the text. Another key principle was our requirement that the 
Yale Daily News be freely available on the Internet. Finally, we wanted 
the newspaper to be fully searchable, browsable, and to include 
advanced search features such as byline and title searches. These 
principals are similar to those elucidated by NDNP and other 
newspaper digitization projects. 

The Initial Decisions 

Our first decisions concerned how we would scan the images and what 
the quality needed to be. Scanning from original source materials will 
almost always provide the truest digital image but depending on the 
format of the source there may be undue complications. The physical 
copies of the Yale Daily News held in the Library's Manuscripts and 
Archives department are tightly bound which hinders how well the 
volumes can be opened. Tightly bound materials pose a few risks for 
digitization projects. Not only can they be damaged during scanning 
but they can also inhibit producing quality digital images. Microfilm is 
easy and inexpensive to digitize but by nature is already a derivative 
format; it will always appear as a black and white photograph of an 
original document. Microfilm can also be of poor quality, out of focus, 
smeared, scratched, or otherwise unreadable. Pictures that have been 
microfilmed often have lost much of their detail. 

Luckily, the Yale Club of New York had a complete copy of the 
Yale Daily News and was willing to donate it to the project. This third 
copy allowed us to avoid the difficult decision of inexpensively 



278 Digitization of the Yale Daily News Historical Archive 

scanning microfilm, scanning bound volumes at a high cost, or 
disbinding our only physical copy, which would balance the cost and 
the quality but would leave us with individual leaves to box and store. 
We were able to disbind, scan and discard these volumes without 
compromising library collections. 

The question of how much storage space would be required for 
the thousands of images we would receive impacted a number of basic 
technical decisions. Bitonal scanning would not represent the 
photographs adequately so grayscale or color images would be 
required. We felt confident that JPG2000 as a reliable file format for 
these kinds of materials. Though the Yale Daily News is an essential 
part of research at Yale, we were not undertaking a scanning project to 
replace the print and microfilm versions of the YDN, but merely to 
provide access. JPG2000 allows large amounts of information to be 
stored in a compressed form, without the compression artifacts and 
other data losses inherent to JPEGs. 

Other universities digitizing complex text and newspapers were 
using METS (The Library of Congress, 2010b) along with ALTO (The 
Library of Congress, 2010a). The METS files enable the program to 
understand the structure of a document, such as a newspaper issue 
with 16 pages. These files were considered "required elements" for our 
projects, as they would enable us to search within the issues. The 
ALTO files provide technical metadata for optical character 
recognition. The combination of the METS files which describe the 
issue of the newspaper with the ALTO files which describe the layout, 
enable full text searching and the highlighting of search words within 
the display of text. The user sees a representative image (a digital 
photograph) of a page. The searchable text created from OCR is a 
stored in a different file, in essence creating a layered document. The 
METS/ALTO links these two such that when a portion or zone of the 
image is highlighted it is associated with the text it represents. Users 
can then search for phrases and see the matching term highlighted on 
the page image. They can also select and copy the text directly from 
the image. 



Digitization of the Yale Daily News Historical Archive 279 



Yale was not in a position to write software to provide access to 
METS/ALTO files for their full functionality. Instead, we decided use 
CONTENTdm, a commercial product that would suit our needs and 
fully utilize the METS/ALTO structure. In order to streamline the 
process as much as possible, we decided to employ vendors to provide 
as many of these services for us as much as possible. The creation of 
the METS/ALTO files were included in our request for proposal (RFP) 
as "Highly Desirable Elements" allowing us to see potential solutions 
our scanning vendors could perform without totally rejecting a 
proposal if a vendor could not provide them. 

In choosing the issues to digitize it may have seemed obvious to 
start with the first issue. However, since the OCD and the Library were 
going to need to fundraise to continue the project, both the Library 
and the OCD felt it would be best to do a range of interesting time 
periods to generate interest with potential donors. In addition, an 
interesting grant funded project that was already underway was 
digitization of World War 1 posters and pamphlets held in the YUL 
collection. We decided to tie into that project and digitize issues from 
1913-1919. In consultation with Manuscripts and Archives it was 
discovered that archivists there use the Yale Daily News frequently in 
answering reference questions and particularly beginning when A. 
Bartlett Giamatti was President of Yale. The records of Yale's 
President, and many university offices, were closed to researchers for 
thirty-five years. The YDN thus provides the best available material 
for historical research on Yale activities between 1978 and 1981. 
Finally, we decided to digitize a very exciting time period that is 
always in demand amongst students, 1967-1970. During this time Yale 
and New Haven experienced student protests, a murder and resulting 
controversial trial of a member of the Black Panthers, and a move 
toward co-education. 

The pilot batch consisting of 800 issues from 13 years, or 
approximately 8000 pages. We were fortunate to be included in a gift 
from the Yale Class of 1945W, which is the class of students from 
1939-1945 whose initial Yale education was interrupted by World War 
II and completed their degree in an accelerated program upon 



280 Digitization of the Yale Daily News Historical Archive 



returning. This provided another 6 years of material bringing the total 
number of pages to 24,000. 

Selection of Vendor and System 

In July 2006 we sent the RFP to four vendors and received responses. 
Most of the vendors were able to fulfill our required and desired 
elements using similar technology. The RFP process and careful 
evaluation of the samples supplied by the vendors was illuminating. 
Through this process, we were able to learn more about the way 
different vendors work and we had a chance to test drive the solutions 
to our problem and evaluate different scanning techniques. 

All of the vendors used DocWorks software to OCR the files, zone 
the articles, and create the METS/ALTO files. Then they prepared the 
package of files for loading into CONTENTdm, using a special loader 
developed by CCS, the company that created DocWorks. We were 
able to provide access to the newspaper by using CONTENTdm. This 
software package, used by many academic libraries, enables users to 
do full text searching on an issue or across all of our newspapers. The 
functionality not only allows viewing images of each page but also 
permits the user to click on an article to view the complete article by 
itself. The recognized text is also accessible, useful for cutting and 
pasting. CONTENTdm can automatically create PDFs of the issue on 
the fly for printing. Finally, because of the zoning and the ALTO files, 
when an article goes on to another page, CONTENTdm pulls all of the 
parts of an article together into one screen, making for easy reading 

We chose to work with Digital Divide Data, a nonprofit company 
based in New York City. Digital Divide Data employs young people in 
third world countries. Their employees not only learn IT skills but also 
attend classes part time in an effort to break the cycle of poverty. 

Digital Divide Data gave us several options for scanning the 
newspapers. Because they are old, the newspapers had a yellowish 
tinge. We chose to scan the newspapers in grayscale, which provides a 
smaller file size and quicker loading of CONTENTdm. We also chose 
to process the text so that the background and text have a high 
contrast, while leaving the images in grayscale. The result is easy to 



Digitization of the Yale Daily News Historical Archive 281 



read, easy and quick to print, and true to the original intent of 
newsprint 

Workflow 

For strategic reasons, YUL does not scan the newspaper issues 
chronologically, making tracking the project complicated but a high 
priority so that the various participants could at any time see the 
status of the project and if any part was held up. A database was 
created for each volume of the newspaper. Fields were added 
including a pull down list for the status of each volume. Because 
various personnel throughout the library were working on the project 
this database helps everyone know what the status is of every volume. 

When volumes have funding designated the status is changed so 
staff knows to prepare those volumes next. The volumes are collated 
and a manifest is created that includes basic information such as the 
volume, number, date, and page count in each issue. Missing volumes, 
printing errors in the enumeration, and possible inhibitions to 
scanning are also recorded. The volumes are then disbound with the 
manifest in hand. Pages missing a significant amount of text are 
removed with the rest of the issue it belongs to. These issues are 
recorded in the database under "Issues Needing Replacement." 
Microfilm will be used to scan missing issues and in order to keep a 
visual consistency entire issues are scanned from the same sources. 
For efficiency sake, missing issues are gathered and microfilm is sent 
and scanned in batches. 

Once a volume has been disbound the manifest is printed and tied 
with the volume. Preparing volumes is frequently performed in 
batches between shipments. This allows us to have a number of 
batches ready and waiting until the shipment is due making deadlines 
easier and less harried. 

Regular shipments are usually sent to Brechin Group, a 
digitization vendor subcontracted by Digital Divide Data, at the 
beginning of each month. After Brechin scans the volumes, the digital 
files are then stored to hard drives and sent to Digital Divide Data's 
Cambodia office for processing where the images are zoned, metadata 



282 Digitization of the Yale Daily News Historical Archive 

is created, and the CONTENTdm files are created. Once processing is 
complete the hard drives are then sent back to Yale where the files are 
upload directly into our CONTENTdm test server for quality control. 

Quality Control 

Given the quantity of items in every shipment, it is not possible to 
check every page. Instead sample issues are checked from every batch 
returned by the vendor using the ANSI/ASQ Z1.4-2003 (American 
Society for Quality, 2003) standard quality control procedures. This 
standard clearly defines how many items need to be sampled, the 
acceptable error rate, and when to accept or reject a batch based on 
the number of errors discovered. This standard was developed based 
on, and is almost identical to, an old military standard (i.e., MIL-STD- 
105E, Department of Defense, 1989) for inspecting shipments. 

This method is based on the idea that there is a level of error one 
is willing to tolerate but if the error rate is too high the entire batch 
will be rejected and reprocessed. The Yale Daily News Project 
identifies four separate areas that need to meet specified quality rates: 
Image Quality, Zoning, Headlines & Authors, and OCR. If a particular 
area fails the vendor only has to redo that particular area, not 
reprocess the entire batch from the beginning. This helps zero in on 
particular problems for our vendors, though it complicates how we 
select the samples. Ideally, the sample would be totally random, but 
this is not feasible considering the samples are based on the particular 
units being checked; so the image quality is based on each page as a 
unit, but the zoning error rate is based on each article as a unit. It isn't 
reasonable to line up all the pages, let alone the articles, and randomly 
sample them considering complete issues are loaded into 
CONTENTdm. To get around this, a system was created to convert 
article and pages into whole issues to be checked. 

A quality control tool was created that helps staff through the 
process that is based off the ANSI standard. The manifests for each 
batch are added and the inspection level is chosen. The tool then 
determines how big the sample sizes are as well as the rejection 
thresholds. A random number generator selects the issues that will 



Digitization of the Yale Daily News Historical Archive 283 



comprise the sample. The sample issues are searched in the 
CONTENTdm test server where the batch was loaded. Errors are 
recorded in the tool as well as notes such that it can be easily found 
again if needed. Once complete the tool calculates the results and tells 
staff appropriate actions. Once a batch is approved it is moved from 
the test system to the live system. 

Copy Right 

When we started the Yale Daily News Historical Archive, the copyright 
situation seemed straightforward. Copyright of the YDN belongs to 
the Yale Daily News Publishing Co., Inc. This company is run by the 
student officers of the YDN, under advisement of a professional 
manager. The project was conducted in close collaboration with the 
officers of YDN and the Oldest College Daily Foundation, and we had 
their permission to digitize and make material freely available. Early 
in the project we concentrated on the earliest material from the 1880. 
This material was mainly text written by students, with some 
advertisements. There were no photographs. Subsequently material 
from World War I was digitized, but material was pre-1923 and again 
consisted mainly of text written by students, although a few 
photographs began to appear. 

It was not until we started work on material from the 1960's that 
we grew concerned that we might have a problem. In the 1960's the 
YDN began to run comic strips, some of which were produced by Yale 
students. Most notably, a strip called Bull Tales first appeared in 
September 1968 written by an undergraduate named Garry Trudeau. 
After Mr. Trudeau graduated he changed the name of the strip to 
Doonesbury which was then syndicated nationally, including in the 
YDN. Peanuts also regularly ran in the YDN throughout the 1960's, 
1970's and later years. The inclusion of these popular and copyright 
protected comic strips raised red flags for us: were we allowed to 
include this material, and did we need to seek permission to do so? 
We worried that we would need to excise this material from the 
digitized copies of the papers. 



284 Digitization of the Yale Daily News Historical Archive 

The comic strips were not unique material that could not be 
found elsewhere, but it did seem that at a time of social and political 
upheaval at Yale and the entire nation, the YDN staff included the 
strips for a purpose, and they did play a part in the tone of the YDN. 
Doonesbury was often overt political comment on current events: 
Peanuts less so, but still was part of a social commentary. The 
digitized YDN would not be complete without the inclusion of the 
strips as they first ran in the paper. We decided that a safe course of 
action would be to contact the rights holders, Trudeau and the estate 
of Charles Shultz. In both cases, permission was given to use this 
material. 

We were lucky that this issue was a problem for other projects as 
well. In June 2008, at the same time we were granted permission by 
the authors to run both strips, the 11th circuit court of Appeals in 
Atlanta rendered a decision in the case of Greenberg v. National 
Geographic Society siding with the NGS. Greenberg had sued NGS for 
including in a digitized version of the magazine material written by 
Greenberg and originally published in the printed magazine. 
Greenberg claimed that NGS only had permission to use his material 
in the original publication, and in reusing his work in the digital 
version had violated his rights under Section 201c of the Copyright 
Act. In siding with NGS, the court found that the use of Greenberg's 
and other's work was permissible because the magazine was faithfully 
reproduced and presented as the original, with material presented in 
its original context. This was in contrast to the 2001 finding in Tasini v 
the New York Times Co., where the Supreme Court found that the 
rights of freelancers were violated because in the digital product in 
question individual articles could by viewed individually, without the 
original context of surrounding material. The difference between these 
decisions ultimately lays in how the digitized content is displayed. The 
publisher has the rights to the issue as a whole and can repackage it as 
a whole, but they cannot split out the parts and make them into a 
different product. In our presentation of the YDN, each issue is 
presented in its entirety as a faithful and full reproduction of the 
original. While searches will indicate individual articles (and 
occasionally comic strips) the user always finds that material in the 



Digitization of the Yale Daily News Historical Archive 285 



context of its original issue. This decision meant that although we did 
get permission from Garry Trudeau and Charles Schultz to use their 
strips it was not required that permission was specifically granted to 
include them. 

Challenges involving Workflow 
and Data Correction 

Anytime a new project starts, especially with a new vendor, there is a 
period of adjustment and problems are expected. Vendors work with 
numerous customers with various expectations. It is extremely 
important for customers to be as clear about their expected outcomes 
and requirements as possible. Good vendors will make every attempt 
to satisfy those requirements. 

In earlier volumes of the Yale Daily News it was not common for 
authors to be named in the articles. During a batch of later years, 
when it became common to name authors, it was discovered that the 
authors' names did not display with other metadata in the header bar 
in the pop up window. This was recorded as an error as it was our 
understanding that fields tagged as "authors" would show up with the 
title of the article. In discussing this with the vendor they were in fact 
tagging them correctly and gave us a short list of issues to check in our 
CONTENTdm installation for why author names were not being 
displayed. 

Other problems were found before the material left the library. As 
we started disbinding older volumes there was some damage to the 
pages. Examples of types of damage were photographed and sent to 
the project staff so decisions could be made on where the cut-off of 
acceptable versus unacceptable damage would be made. It was 
ultimately decided that some minor loss of text was acceptable as long 
as the user could still surmise the lost text and it was not in an area 
that may be in a targeted search, such as the title of the article. 

Other challenges have come up during the course of the project 
but because of our good relationship with the vendor we can work out 
solutions easily. 



286 Digitization of the Yale Daily News Historical Archive 

Funding and Sustainability 

The project was originally funded with contributions from the YUL 
and the Oldest College Daily Foundation. In addition the YUL 
matched its original funding amount for one additional year. At the 
same time the Library Development Office added the YDN digitization 
project to its priority list and began actively seeking contributions. The 
Yale Class of 1945W (the World War II years) signed on quickly to 
fund the eight academic years covering 1940-1948. 

As we moved the project from the pilot phase into production we 
faced challenges brought about by the depressed economic climate of 
early 2009. This became the most important factor in determining 
how the project would continue. Decreases in the Library budget 
resulted in the elimination of additional library funds. Therefore, 
digitization of additional content is currently being funded entirely by 
donor contributions. 

The Library's Development Office continues to prioritize this 
project on its development list and actively works with potential 
donors to identify time periods that may be of individual interest and 
are available for funding. Contributions are publicly acknowledged on 
the Library's website (http://images.library.yale.edu/ 
digitalcollections / ydnAcknowledgments . aspx) . 

As we move forward issues for digitization will be selected based 
on a variety of criteria. Donor funding may be given for specific years. 
These years will be given the highest priority in the digitization queue. 
Next, if donors do not select specific years for digitization, priority will 
be given to years that are in demand by researchers at Yale. Specific 
years or eras are requested repeatedly - including the 1960's. 
Identification, digitization and availability via CONTENTdm of these 
specific years can improve services to researchers by providing them 
immediate access to the information they require while at the same 
time increasing staff productivity by avoiding repeated copying of the 
same articles. Once we have digitized all content identified as highly 
desirable by researchers we will fill in remaining year gaps beginning 
with the oldest content. 



Digitization of the Yale Daily News Historical Archive 287 



Conclusion 

The Yale Daily News digitalization has been a challenging and 
rewarding project. It has been a partnership with organizations 
outside of the library, such as the OCD Foundation, and also a great 
opportunity within the library for various departments to work 
together. The core team within the library comes from five separate 
departments: Cataloging and Metadata Services, Electronic 
Collections, Library Access Integrated Services, Preservation, and 
Usability & Assessment. This group meets regularly and work out 
issues that each department have interests in finding collaborative 
solutions, many of which were able to move into other fledgling 
projects. In addition, Preservation department staff performs the 
tasks of preparing the newspaper for digitization and for checking the 
quality of the digital files as they return. This process utilized skills the 
department already had, expanding them for newspaper digitization. 

On the fundraising side it has proven to be a great springboard. 
Yale graduates tend to be quite loyal and they have frequently been 
very interested in looking back into the digitized YDN content that has 
been created. Not only are they giving back to help complete the 
project but it has also been an eye catching project where other 
potential projects can be discussed and funded. 

Finally, digitizing the Yale Daily News has enabled the library to 
produce a free, highly useful, and unique digitized resource for both 
Yale University and other researchers. The expertise gained in this 
project has enabled YUL staff to build on this success with other, more 
complex, digital projects. YUL staff can now successfully digitize 
varied material such as maps, annotated manuscripts and books. Mass 
digitization of books may free libraries to concentrate on high quality 
digitization of unique material that is present in nearly all public and 
academic libraries. A newspaper digitization project can be an 
excellent springboard for digitization of this diverse set of material. 



288 Digitization of the Yale Daily News Historical Archive 

References 

American Society for Quality. (2003). ANSI/ASQ Zi. 4-2003: 
Sampling procedures and tables for inspection by attributes. 
Milwaukee: American Society for Quality. 

Department of Defense. (1989). Military standard. Sampling 
procedures and tables for inspection by attributes. Washington, 
D.C.: Department of Defense. 

The Library of Congress. (2009). National digital newspaper 

program. Retrieve May 20, 2010 from http://www.loc.gov/ndnp/ 

The Library of Congress. (2010a). ALTO: Technical metadata for 
optical character recognition. Retrieve April 10, 2010 from 
http : / / www.loc.gov/ standards / alto/ . 

The Library of Congress. (2010b). METS Metadata encoding & 
transmission standard. Retrieve April 10, 2010 from 
http : / / www.loc.gov/ standards / mets / 



New Jersey Digital Legal Library 



Wei Fang (Rutgers University) 



Abstract 

The New Jersey Digital Legal Library (NJDLL) was launched in 2003 
and it's been up and running since then. NJDLL has sixteen 
collections with about 500 video clips and 20,000 electronic 
documents. It has served over 300,000 visitors and is popular among 
legal professionals and researchers. The goal of this project is to create 
a Web-based digital library where patrons can browse and search for 
previously unavailable New Jersey legal information. In this article, 
the author presents the background history of the NJDLL, how it was 
constructed, workflows, and how problems were solved in depth. 
Future development suggestions are also discussed. 

Keywords: digital library, legal library, XML, LAMP, repository, 
Supreme Court, oral arguments. 



Introduction 

As part of the Rutgers University School of Law, the Rutgers 
University Law Library - Newark with its more than half a million 
volumes of printed books is the largest law library in New Jersey. Its 
collections include state and federal statutes and court decisions, 
federal and New Jersey regulations and administrative decisions, 
federal and New Jersey legislative history materials, the codes of 
ordinances for many New Jersey municipalities, Anglo-American legal 
periodicals, the primary materials of international law, extensive 
historical materials on English law, and a special collection of 



289 



290 New Jersey Digital Legal Library 



criminology and criminal justice materials. The library has served the 
law school and the surrounding legal community for more than a 
century. Our law librarians wanted to share the unique collections that 
the law library had with others in the world and the only way to do it 
was to convert the collections into digital documents and host them 
online via the Internet. 

The NJDLL project focuses on legal materials related to New 
Jersey. The target audience groups are legal professionals, historians, 
researchers and the general public. After more than five years of 
development, its collections include electronic legal documents as well 
as New Jersey Supreme Court oral arguments archives. The NJDLL 
project was first proposed in the fall of 2000 by Paul Axel-Lute, the 
collection development librarian, and was initialized in spring of 2002 
with hardware and software funded by Rutgers University. In April 
2003, the digital library was inaugurated with three collections: 1) 
New Jersey Administrative Reports 2) New Jersey Executive Orders 
and 3) New Jersey Attorney General Opinions. In 2005, the law 
library negotiated with the Supreme Court of New Jersey to be the 
host of its Oral Argument video clips using Microsoft Windows Media 
Video technology. We are also hosting the digital collection of the New 
Jersey Council On Affordable Housing's decisions. The NJDLL has 
now grown to become a popular digital library with sixteen collections 
ranging from year 1779 to present (see table below). 



Collection 


Description 


Holmes v. Walton 


Case file transcriptions (1779-1783) etc. 


New Jersey Administrative Reports 


First Series; Covers 1979-1991 


New Jersey Attorney General 
Opinions 


Covers 1949-1998 


New Jersey Council on Affordable 
Housing (COAH) 


Decisions, 1986-Present 


New Jersey Department of Human 
Services, Administrative Orders 


Orders from 1 977 forward 


New Jersey Executive Orders 


Covers 1941 - January 1990 


New Jersey Governor's Code of 
Conduct 


Text of Code 



New Jersey Digital Legal Library 291 



New Jersey Affordable Housing 
Archive 


Various resources relating to affordable 
housing in New Jersey 


New Jersey Legislative Counsel 
Opinions 


Opinions Since 2004 


Supreme Court of New Jersey Oral 
Arguments 


Video Archive of Arguments since 2005 


The Law of Slavery in New Jersey 


Bibliography, Statutes and Cases 


The Mount Laurel Material of The 
Affordable Housing Archive, New 
Jersey 


History and Selected Documents 


Miscellaneous Documents 


Topical Legal Documents 


Mercer Beasley & Univ. of Newark 
Law Reviews 


Digitized versions of select Law Reviews 


Statute Compilations 


Compilations of New Jersey Statutes 


Weintraub Lecture Series 


Selected lectures from the law school's 
annual lecture series. 



NJDLL Collection List 



System Architecture 

After the project was started, a new position was created and the 
library hired Kevin Reiss in October 2003 as the digital services 
librarian who served the library till March 2006. He played a major 
role in setting up the overall architecture of the site. Meanwhile, the 
library purchased a new server that has two Intel Xeon processors, 
two 74 GB hard drive in RAID 1 configuration, two 100 MB network 
interface cards and 8 GB internal memory. The server was set up in an 
air-conditioned, secure room with automatic fire extinguisher. It was 
then connected to a high-speed network for the best user experience. 
The library decided to use as much open source software as possible. 
The best part of the open source software is that it is free. Also, unlike 
most of the commercial software, users can look through all the source 
codes to make sure it is safe to use if necessary. As a result, the entire 
project was built with Redhat Linux, Apache server, MySQL database 
server, PHP and XML parser. We also used Swish-e search engine 
(Swish-e, 2009) for the searching functionality. Operating system, 



292 New Jersey Digital Legal Library 



server software and supporting software were installed within a 
month. A tape-based system performs full backs up of the server once 
a week. 

Usability and Accessibility 

The library administration identified three collections to get the 
project started. After law librarians selected documents in these three 
collections and the server was tested, the digital services department 
of the law library received go-ahead signal to start building the site. 
The first and critical step was to determine the digital documents' 
formats. There were many digital document formats that we could 
choose from, including Adobe PDF, Microsoft Word, Corel 
WordPerfect, etc. After some careful evaluations and tests, we chose 
Adobe PDF and DjVu as the formats for our digital documents of the 
collections. Adobe PDF and DjVu are widely supported by the open 
source community and they both have free readers available for PC 
and Mac users to download. For all scanned materials, we defined a 
protocol that requires all the documents to be saved in 300 DPI 
monochrome in both TIF format and JPEG format. We also 
performed Optical Characters Recognition (OCR) on scanned 
documents. Depending on the size of the collection, the library 
outsourced boxes of documents to an outside digitization company. 
The library would contact a digitization company and provide 
specifications for the company to follow. For smaller sized and 
consistently updated collections, instead of outsourcing, the library 
purchased a high speed sheet scanner and an overhead book scanner 
to scan all the documents in house, since this method would cost less 
and provide the most convenience in the long run. After converting 
paper-based documents into digital documents Adobe PDF, Djvu, and 
plain text, XML files for the metadata of the documents were 
composed manually. 

The site was built using PHP language with item information 
stored into a MySQL database. A menu system was designed and it 
used XML file to generate the menu on the fly. For instance, when a 
patron visits the site, the PHP language will produce the page layout 
and extract item information from the database while the XML parser 



New Jersey Digital Legal Library 293 



generates the menu for that particular page. This provides the system 
manager with flexibility and accessibility to manage the site. 

User interface could be the most critical element when designing 
a Web site. Digital libraries have traditionally been regarded as 
difficult to use (Borgman, 2000). When the site was constructed in 
2003, most people used computers with dial-up connection and low 
resolution monitors. The library, therefore, designed the NJDLL with 
a fixed resolution at 800 pixels in width. To reduce the network load 
and provide a faster speed, the site used only one graphic, the New 
Jersey logo, on most of its pages. We also designed a simple interface 
with an interactive menu list - just to make the uers to be comfortable 
with (Markri S., etc al., 2006). 

Workflows 

As described in the System Architecture section, the library uses 
as many open source tools as possible. However, this does not mean 
that the library objects to using commercial software. We use Adobe 
Acrobat Professional (Adobe Acrobat Professional, 2009) to generate 
PDF files and use Any2 Djvu Server (Any2DjVu Server, 2002) to 
generate DjVu files from PDF files. The Adobe Acrobat costs about 
one hundred dollars per university license and the Any2 DjVu service 
is free of charge. Similarly, work study students use commercial 
software Abbyy Finereader to perform OCR tasks and use open source 
software, Notepad++ (Notepad++, 2009), to compose metadata files. 
For video files, work study student use Windows Movie Maker from 
Microsoft (WMM, 2004) to convert video clips. Although the WMM is 
not open source, it is freely available. 

Throughout the years, the library has developed and improved 
three protocols to process digital objects based on the types of the 
resources as well as the scale of the resources (Figure RU-i). 

Internal Resources 

The library defines internal resources as collections over which 
the library has total control. Reference librarians are in charge of 
identifying each and every item to be scanned. Depending on the size 
of the collection, the library administrators decide whether to scan the 



294 New Jersey Digital Legal Library 



documents in house or to outsource them. When a collection is 
scanned in house, work study students, hired from the law school and 
other departments of the university, are assigned to scan the 
documents using the high-speed sheet scanner for loose leaf pages 
and using overhead book scanners for rare documents. They also 
generate all metadata files associated with each item. When a 
collection is outsourced, the library contacts the outside vendor to 
provide digitized documents using CDs or DVDs. The digital services 
librarian will do a quality check and verify the quantity. If everything 
meets the requirements, work study students will be assigned to 
compose metadata for the documents. 



Digital Objects 




Printed documents 
in large quantity 




r Digitize doc u merits ^ ~"\ /r 
and OCRusing 
predefined 
specifications 




Post new content 



Define 
specifications 



Sfiipcmt documents 
to print shops 

• 1 

Guartityand quality 
checkof received 
digital documents 



3 



Compose Metadata 



Post new content 



Figure RU-1: Workflows 



Convert video clips 
using predefined 
spedficat 




Compose metadata 
files for each video 
clip 

T" 




When the content is ready to be updated, the digital services 
librarian uploads metadata and digitized documents to the server and 
runs a UNIX script to generate database records and item information 



New Jersey Digital Legal Library 295 



based on the metadata file. Librarians and IT personnel use SSH to 
upload files and run UNIX commands. To simplify daily work, they 
also use Webmin and phpAdmin to manage the site and the database. 

External Resources 

External resources are the resources that the library has to 
acquire from outside of the library and the library has no or very 
limited control over the quality of the resources. The NJDLL has two 
collections that use external resources: New Jersey Supreme Court 
Oral Arguments (NJSCOA) and legal documents from New Jersey 
Council On Affordable Housing (COAH). 

The library uses the in-house protocol to scan materials for the 
COAH collections. For the NJSCOA, the library assigns work study 
students who are law school students for their capability of 
interpreting legal materials. Work study students are trained to 
produce metadata and video clips for this collection based on the files 
received from the IT office of the New Jersey Supreme Court. As a 
procedure jointly developed by the Court and the library, after the 
case is argued, the IT office will transfer video recordings in standard 
broadcast resolution along with transcripts and detailed argument 
information to a FTP server. When we decided to host the NJSCOA 
video archives, we developed a new database and procedure to convert 
the original high resolution video clips received from the Court into 
low-bandwidth and medium-bandwidth video clips. For each and 
every video recording, work study students will produce two video 
clips for medium and low bandwidth, metadata in XML format. After 
a month of the argument date, cases will be uploaded to the server and 
posted by the digital services librarian using a UNIX script. 

Quality Control 

Quality is a very important element of any successful Web site. As 
the matter of fact, if one site does not provide a good quality expected 
by the visitors, it will soon lose credit, attract no visitor and dry out 
quickly. Using existing standards is very important in quality control, 
product usability and documents preservation. Work students will 



296 New Jersey Digital Legal Library 



encode metadata records for each digital record by following the 
standards listed below: 

1) METS: Metadata Encoding and Transmission Standard 
(METS, 2009) 

2) EAD: Encoded Archival Description (EAD, 2002) 

3) Text will follow the Text Encoding Initiative (TEI) P5 guideline 
(P5, 2009) 

New work study students need to be trained before they can work 
on any assignment. The library has composed many procedures for 
the students to follow. Before the metadata records can be used to 
generate web pages and shared online, the digital services librarian 
will check them for quality. 

Backup Strategy 

A modern library should provide two basic functions: preserve 
and serve. As important as serving knowledge to the people, our 
library also has long history of preservation. All master scan images in 
NJDLL project are stored in uncompressed TIFF format. Metadata 
files are composed using XML formats, and we have a well-oiled 
backup strategy to ensure that we don't lose these important digital 
assets when disaster strikes. 

A mirrored RAID 1 hard drive array is used as the first tier of 
backup. The IT office controls an external tape drive that backs up the 
entire server every week. To save tapes, the IT office recycles tapes 
after two months of the last backup. The digital services department 
also backs up the server remotely by using a server-grade external 
hard drives as the third tier of the backup. Each tier holds a full set of 
backups, including normal backup sets and incremental backup sets. 

Frequency of backups: One normal backup would be performed 
monthly. Every week, an incremental backup would be performed. 
Two physical locations are assigned for the first tier and third tier 
backup devices. 



New Jersey Digital Legal Library 297 



Problems Encountered and Counter Measurements 

During the past five years, the biggest technical problem we had 
was security. We started the Web site and made it public in 2003. The 
site did not have any protection initially since we thought that we are 
an organizational site and it is on a state university's network. The site 
was hacked within a month. After the incident, the server and the site 
were restored and we installed firewall and trapwire on the server. The 
site was also readjusted to accept administrative connections from 
Rutgers University IP addresses only. Both the IT office of the law 
school and the library monitor the activities on the site consistently 
and update server patches as soon as they are available. 

Structurally, the site was built by grouping small collections one 
at a time. Each collection had its uniqueness in terms of materials and 
focus scope. The site had to be fine tuned each time there was a new 
collection to be added. When we added the Supreme Court Oral 
Argument into our collection, one big issue was that we had to use the 
Court file name schema that is difficult to use and understand by our 
work study students. For instance, a dash and an underscore may 
make no difference to the student workers, but the UNIX server will 
treat them differently. We had to communicate with the Court and 
adjust the naming schema between the library and the Court. 

Another thing which may not be as obvious as the name schema 
was the technical documents that the site management team has to 
compose. When the server was installed and the site was built, 
everyone involved focused on their work and paid little attention to 
writing down technical notes and documents. When some of the staff 
members left for other positions, the successors had a lot of difficulty 
in understanding why the script or the system was designed in its way. 
We now require all staff members to write technical documents on a 
regular basis and new staff members have to read those documents 
before performing any work on the server or the system. 

The technologies we have used are becoming obsolete gradually. 
One example is that the library has been using the Windows Media 
Video for years and keeps getting phone calls from Mac users 
complaining that they are not able to view the contents. Considering 



298 New Jersey Digital Legal Library 

the fact that more and more students and attorneys are using Mac 
computers, this is something the library has to address. We are 
building our own Youtube-like video repository to host copyrighted 
video recording (Fang, 2009) and we may use the same video 
repository to host the Supreme Court oral arguments in the future. 

One issue that is hard to get around is that the inconsistent 
quality of work study students. Every semester the library has to hire 
students to scan documents, compose metadata, and convert video 
clips. As required by the Rutgers University, the library can only 
choose students within the university. Students may come from the 
law school or any other department of the university. Generally, 
undergraduate students perform tasks with the lowest quality. They 
have no idea about legal documents and make errors all the time. On 
the other hand, law school students perform all the tasks the best. 
They get their hands on quickly and enthusiastically. However, they 
are always on a busy schedule especially during the exam session. In 
order to solve this issue, the library composes protocols and asks new 
work study students to memorize them. This can reduce errors in 
some degree though it cannot completely eliminate them. 

The project has been up-and-running for more than five years. 
We would like to track website activities. Apache web server provides 
a log file and the server has software that can analyze the log file. 
However, it does not provide many details on each page or file viewed 
by the visitors. We tried other new methods on the site, including Web 
analytic software and some online Web analytics providers. Not all 
these methods were reliable. In one particular case, the Web site 
stopped responding since it was waiting to load tracking codes from 
the tracking provider. After deploying Google Analytics (Google, 
2009) on the library's main Web site, we got positive results (Fang, 
2007). The library finally decided to use Google Analytics on the 
NJDLL site for its functionality and reputation. Now, after two and 
half years of using the Google Analytics to track our site, we are 
satisfied with the results. 



New Jersey Digital Legal Library 299 



Important skills learned from doing this project 

1. Security always has the top priority: the server should be behind a 
hardware or software firewall and set up in a way that 
administrators can only log in from the Intranet. Just in case 
something bad happens, a fully tested backup plan and 
equipments should be used. 

2. Use uniform name schema: when collaborating with another 
organization to develop a new collection, it is extremely important 
to go through naming schemas before starting coding. Every detail 
has to be thought about thoroughly or later on the smallest hole 
can grow unmanageable. 

3. Document everything: Spending five minutes writing a short note 
may save you a day of work later. 

4. The last day of old technology is better than the first day of the 
new technology: Five years may seem to be a short period in the 
technology field, but we've seen a lot of new hardware and 
software faded away. Our take on the technology is that we stick 
with a time-proofing reliable platform as long as it meets our 
requirements and visitors' expectations. 

Future development 

As of now, we have about 500+ video clips and about 20,000 
electronic documents on the site. The site attracts about 40,000 
visitors annually, which is just enough for our hardware to handle. 
The law library will keep maintaining and adding new items to the site 
every semester. New collections will also be added once they are 
digitized. The biggest goal for the next few years would be merging our 
video content from Windows Media Player format to the more popular 
Adobe Flash Video format. 

Currently the entire site is being indexed by the Swish-e search 
engine on a weekly basis. We have some good experiences with this 
free search engine, but it is hard to be reconfigured and needs some 
workaround procedures to get the XML files indexed correctly. We are 
planning to replace the Swish-e search engine with Google's Custom 
Search, with which the university has contract. 



300 New Jersey Digital Legal Library 



When the time comes to replace our current server's software 
platform, we will use Ubuntu server edition (Ubuntu, 2009) and 
Drupal (Drupal, 2009) as the new platform and merge over all the 
existing contents. Ubuntu server edition is a Linux based LAMP ready 
operating system. Drupal is an open source content management 
system, which has been used by thousands of Web sites, including 
government Web sites such as the whitehouse.gov. 

References 

Adobe Acrobat Professional. (2009). Adobe and PDF. Retrieved from 
http : / / www. adobe, com/ products/ acrobat/ adobepdf.html 

Any2 DjVu Server. (2002). Welcome to the DjVu Any2Djvu Server. 
Retrieved from http://any2djvu.djvuz0ne.org/index.html 

Borgman, C. (2000). From Gutenberg to the global information 
infrastructure: access to information in the networked world. 
Cambridge, MA, USA. MIT Press. 

Drupal. (2009). About Drupal. Retrieved from 
http:/ / drupal.org/ about 

EAD. (2002). Encoded archival description version 2002. Retrieved 
from http://www.loc.gov/ead/ 

Fang, W. (2007). Using Google analytics for improving library 
website content and design: A case study. In Library Philosophy 
and Practice, Annual Vol. 2007 (Special Issue Libraries and 
Google) Retrieved from http://www.webpages.uidaho.edu/ 
-mbolin/ fang.pdf 

Fang, W. (2009). Online law school video repository: The flash way. 
Computers in Libraries, Vol. 29, No. 6. 

Google. (2009). About Google Analytics. Retrieved from 
http://www.google.com/analytics/ 

Makri, S., Blandford, A., & Cox, A. (2006). Studying law students' 
information seeking behaviour to inform the design of digital law 
libraries. Paper presented at the Alicante, Spain. 



New Jersey Digital Legal Library 301 



METS. (2007). METS schema, & documentation. Retrieved from 
http : / / www.loc.gov/ standards / mets / mets-schemadocs.html 

Notepad++. ( 2009). About Notepad++. Retrieved from 
http:/ / notepad-plus.sourceforge.net/ uk/ site.htm 

P5. (2009). TEI: P5 guidelines. Retrieved from http://www.tei- 
c. org/ Guidelines/P5 / 

Swish-e. (2009). About Swish-e. Retrieved from http://swish-e.org/ 

Ubuntu. (2009). What is Ubuntu. Retrieved from 
http://www.ubuntu.com/products/whatisubuntu 

WMM. (2004). Windows Movie Maker 2.1 Download. Retrieved from 
http : / / www.microsoft.com / windowsxp / downloads/ 
updates/moviemaker2.mspx 



Pratt Institute: A Historical Snapshot of 
Campus and Area 



Paul Schlotthauer (Pratt Institute Libraries) 



Abstract 

The chapter discusses the planning and procedures for two 
digitization grants received from METRO in 2005 and 2006 by the 
Pratt Institute Libraries. Despite the creation of a timeline, 
unexpected issues necessitated adjustments. Selection of materials 
was more labor intensive and time consuming than expected. Our 
collection consisted of different formats, which required managing 
significant differences in metadata content and interpretation. OAI- 
compliancy was problematic with our image management software. 
Also, we had modified the Dublin Core fields, which created 
compatibility issues with OAI gateways. In 2006, we planned for 
potential pitfalls based on our previous experiences, but still 
encountered delays and problems, such as color management. In 
2008 we mounted our images on Flickr and saw a marked increase in 
use. 

Keywords: Archives, Best practices, Bookplate, Color correction, 
Compliancy, Digitization, Historical images, Image management 
software, Imaging, Intellectual control, Metadata Template, Negative, 
Pratt Institute, Preservation, Project management, Selection criteria, 
Selection guidelines, Special collections. 



302 



Pratt Institute: A Historical Snapshot of Campus and Area 303 

Introduction 

In 2004, when we applied for the METRO grant for digitization, the 
library at Pratt Institute's Brooklyn campus had a Visual Resources 
Center that, since 1997, had been digitizing its collection of over 
160,000 slides for teaching and research purposes. We had staff and 
student workers with training and experience in digitization, as well as 
dedicated space, equipment, and technical support. Our digitized 
images were available online, so we had some acquaintance with the 
issues surrounding online collections. 

For us, therefore, it was not so much a question of beginning a 
digitization initiative as it was beginning a new phase in our already- 
existing digitization program. Our hardware was no longer state of the 
art. The evolution of metadata schemas and best practices, such as 
Western States Digital Imaging Best Practices (later revised by the 
Bibliographic Center for Research as BCR's CDP Digital Imaging Best 
Practices) (Bibliographic Center for Research [BCR], 2008), meant 
that our own protocols were inconsistent if not out of date: they 
worked for us, but they were not truly in step with the most current 
professional standards. We envisioned a METRO grant as an 
opportunity to update and tighten our practices, and believed that the 
resulting increase in efficiency would generate momentum to expand 
our digitization program in order to better meet the increasing 
expectations of our users. 

Determining the Team and the Project 

When we applied for the grant, we began to assemble the team. 
For project manager our Director of Libraries selected our Visual 
Resources Curator, who had managed Pratt's digitization program for 
over a year and had a background in photography, metadata, and 
technology. As the Institute's archivist, I had devoted considerable 
effort to researching, organizing, and preserving the archives' large 
image collection, and was therefore asked to serve as the selector and 
conservator of the materials to be digitized. 

Throughout the course of the project, both the project manager 
and I took several of the digitization training courses offered by 



304 Pratt Institute: A Historical Snapshot of Campus and Area 

METRO, including Digital Imaging, Digital Project Infrastructure, 
Preserving Digital Materials, Metadata for Digital Collections, 
Introduction to XML, Basic Copyright Issues, and Using Adobe 
Photoshop for Image Quality Control. These classes provided us with a 
firm foundation in various aspects of digitization that enabled the 
project to proceed more efficiently. 

For our scanning technician we selected a graduate student from 
the School of Information and Library Science who possessed 
considerable knowledge of imaging, metadata, and Web content 
management issues, and who had prepared, scanned, retouched, and 
cataloged many images in the Visual Resources Center. We also began 
to draft a job description for a project cataloger. 

In determining the experience and training needs of the project 
team, we adhered to the guidelines of the North Carolina Exploring 
Cultural Heritage Online (ECHO) initiative and addressed four 
principal issues: conservation, digitization/encoding, 
metadata/cataloging, and technical development/support (NC ECHO, 
2007). We also planned to use a graduate student from Pratt's 
Communications Design Department to develop the Website for this 
project. As with our scanning technician, we involved both our School 
of Information and Library Science and our School of Art and Design 
in a truly collaborative effort, which gave weight to our assertion that 
digitization was relevant to the entire Institute, not just the Libraries. 

As the image curator, I was faced with the problem of what to 
select from the archives and Special Collections for digitization, given 
the embarrassment of riches at the Institute and METRO'S limit of 
five hundred images per institution. As Janet Gertz (2007) has stated, 
"In selecting well, institutions of all sizes and types concentrate on the 
parts of their collections that are best suited to digitization, make the 
most effective use of the technology, and meet their clients' needs. . . . 
Good selection decisions come through carefully assessing the 
physical nature and content of the original materials, the intellectual 
property rights connected with them, and the requirements for a 
technically sound, well-described, and cost-effective product that 
serves both users' need for access to the content and the institution's 



Pratt Institute: A Historical Snapshot of Campus and Area 305 

need to preserve the materials" (Introduction, paras. 1 and 2). I 
therefore decided to make an eclectic selection from several 
collections rather than focusing on one format or theme. 

I based my broad criteria on subject matter and quality of image. 
My goal was to illustrate the breadth and diversity of the collections 
while emphasizing various facets of the Institute as an academic 
institution, a presence in the local community, and an important 
influence in the fields of art, architecture, fashion, design, and 
education. Our archival image collection existed in a variety of 
formats, including photographs, slides, and negatives, and focused 
primarily on such aspects as the architecture and grounds of the 
campus and the Clinton Hill section of Brooklyn in which it is located; 
student work in art, architecture, fashion, and design; and campus life 
and activities. I selected a number of historical photographs of Pratt's 
buildings, some of which date from the nineteenth century and are 
well known for their architectural features. 

The campus, a twenty-five-acre oasis in the historic district of 
Clinton Hill in northern Brooklyn, was also represented, in part 
because early images of the campus afford views of surrounding 
Brooklyn streets and include long-demolished buildings and 
structures. There were also photographs of classes in session and 
students at work dating from the early decades of Pratt's history. 

These were significant from the standpoint of the history of 
education. They depicted the state-of-the-art studios and equipment 
used for classes in such fields as drawing, woodworking, engineering, 
and physical education. Also important from a socioeconomic 
standpoint was the presence of women and minorities in many of 
these photographs, demonstrating the Institute's progressive outlook 
in an era when racial and gender discrimination were the norm in the 
United States. I included images of the founder and his family, as well 
as some of the mansions they built, as comprising an important part 
of Brooklyn's history. Finally, I chose examples of student work, such 
as architectural models and drawings, designs for clothing and 
interior decoration, sculpture, prints, and posters, which illustrated 
the outstanding body of work produced by Pratt students during its 



306 Pratt Institute: A Historical Snapshot of Campus and Area 

existence. Many of the examples in this category came from negatives 
dating from the 1950s through 1970s, and all gave evidence of the 
creativity fostered by the Institute's faculty. 

From Special Collections I selected prints of hand-colored 
stenciled fashion plates from La Gazette du Bon Ton, considered the 
most influential French fashion magazine of the early twentieth 
century, as well as representatives of Pratt's collection of nineteenth- 
and twentieth-century bookplates. The fashion plates could serve as 
resources for fashion students and designers, while the bookplates 
were outstanding examples of period book art and typography. 

We were gratified to receive a one-year grant from METRO to 
digitize this selection. We first created a timeline incorporating all the 
major activities, which allowed us to track our progress and, at the 
mid-point, enabled us to make the adjustments necessary to complete 
it in time (Figure PRAT-i). 

Imaging took place on an iMacG5 (20-inch, 1.8 GHz, 1 GB RAM, 
250 GB hard drive), an Epson Expression 10000XL scanner, and 
Adobe Photoshop Creative Suite software, all made possible through 
the METRO grant. The raw and enhanced TIFF files were saved to 
Gold Mitsui CD-Rs and a 1.6 TB LaCie external hard drive (the LaCie 
drive not only allowed secondary storage and backup but at times 
became a shared drive for the various computers used for this 
project). The resolution had to be determined by the format: the 
photographs, bookplates, and fashion plates were scanned at 600 dpi, 
whereas the negatives were scanned at 1,200 dpi. 

Once we received the grant, and while the initial scanning was 
taking place, the project manager, scanning technician, and I formed a 
search committee to recruit a digital archive cataloger. This person's 
ability to capture the descriptive and technical metadata would be a 
crucial element in the success of this project, so we insisted that 
candidates possess experience and knowledge of MARC21, AACR2, 
Library of Congress Subject Headings, Dublin Core, and emerging 
descriptive metadata standards. 



Pratt Institute: A Historical Snapshot of Campus and Area 307 



Promotions 


Website 
Design 


Records 
Catalogued 


Images 

Scanned/ 

Edited 










0 
0 


02/05 








to 

8 


| 03/05 








OJ 

0 
0 


1 04/05 






o\ 
0 


& 
0 
0 


05/05 






LA 
O 


LA 
0 
0 


06/05 






UJ 
O 
O 


Edit 


07/05 




Start 
Design 


O 
O 


Finish 


08/05 ! 




Refine 
Design 


LA 
O 
O 




09/05 


Create 


Create 
Site 


Edit 




10/05 


Create 


Launch 


Finish 




1 1/05 


Send 








12/05 



s 

5" 
m 

(X 



3 



Promotions 


Website 
Design 


Records 
Catalogued 


Images 

Scanned/ 

Edited 










0 
0 


02/05 






0 


0 
0 


03/05 






00 
0 


Ui 

0 
0 


04/05 




Begin 
Plans 


o\ 
0 


0 
0 


05/05 




Start 
Design 


0 


LA 
O 
O 


06/05 




Refine 
Design 


u> 
to 
0 


Edit 


07/05 




Create 
Site 


*. 

0 
0 


Finish 


08/05 




Create 
Site 


00 
0 




09/05 


Create 


Create 
Site 


Edit 




10/05 


Create 


Launch 


Finish 




11/05 


Send 








12/05 



Figure PRAT-1 . Schedule for the 2005 Grant. 



308 Pratt Institute: A Historical Snapshot of Campus and Area 

We also expected knowledge of authority control in the online 
environment, as well as experience using bibliographic utilities and 
assigning metadata to digitized materials for Web-based display and 
searching. During the interviews, in addition to asking the usual 
questions, we gave the candidates several images and asked for on- 
the-spot descriptions; this was an effective means (with considerably 
varied results) of determining the candidates' observation powers and 
the kind of descriptive information they would be likely to provide. 
The search committee's diligence was rewarded with a candidate who 
produced outstanding work. 

Descriptive metadata was to follow the Dublin Core elements of 
title, creator, subject, description, date digital, date original, format, 
digitization specifications, resource identifier, rights management, 
contributing institution, publisher, contributor, type, source, 
language, relation, and coverage. Technical metadata was to include 
resolution, size in Kilobytes, height and width in pixels, bit depth, and 
color space. Library of Congress Subject Headings, as well as LC's 
Thesaurus for Graphic Materials, were to provide the controlled 
vocabulary for cataloging content. 

Challenges 

Scheduling 

Despite our careful planning, we encountered a variety of 
challenges, several of which affected our schedule and required 
adjustments. Perhaps the greatest problem was maintaining a 
consistent workflow. All the team members except for the cataloger 
took on the project in addition to their usual responsibilities, which 
impeded the conducting of regular meetings and delayed weekly or 
monthly goals. Because we were spread rather thin at times, it was 
difficult to maintain good communication within the group. We all 
recognized how quickly we would lose control if the team were 
disjointed, so we made special efforts to touch base regularly and keep 
each other informed through e-mail exchanges and periodic meetings. 
Thus, when the project director left before the project was completed, 
it was relatively easy for me to step in and take her place. 



Pratt Institute: A Historical Snapshot of Campus and Area 309 

Selection and Scanning 

Because we did not focus on a specific theme or collection, I 
needed to employ multiple criteria in my selection process, including 
artistic quality, historical significance, and research or special-interest 
potential, to assemble a collection both unique and engaging on 
various levels. I tried to operate within the parameters set forth in 
several resources on this issue (Gertz, 2007; Hazen, Horrell & Merrill- 
Oldham, 1998; Southeastern NY Library Resources Council 
[SENYLRC], 2004), but even so, freedom of choice made it difficult to 
adhere to the maximum number of images mandated by METRO 
without omitting something I thought especially interesting, and in 
fact I had to pare down the number of selections several times. This 
was more labor intensive and took up more time than anticipated, and 
thus a certain amount of haste was required given the time frame. 
Even so, I was late in delivering the final selection. For future projects, 
we knew we would have to begin the selection process sooner and 
allow more time for its completion. Of course, focusing on one theme 
or collection would also ease the problem. 

Another important reason for the delay was the fact that only the 
negatives had pre-existing unique identification numbers. The 
bookplates, fashion plates, and most of the photographshad no 
identifiers at all. We had to devise a system for these at the time of 
selection. The simple act of numbering all these images, especially for 
those in fragile condition, was extraordinarily laborious and time 
consuming, even with the aid of student workers, yet it would have 
been disastrous not to have done it. Given the time constraints for this 
project we had no choice but to plow ahead, but, based on this 
experience, I can certainly endorse the view that, whenever possible, 
one should avoid digitizing materials lacking a minimal level of 
intellectual control. 

The project's wide range of materials also made the development 
of a systemized workflow and management of the digital results 
difficult. The scanning process required a tremendous amount of 
patience, especially with the negatives, for which the time needed to 
correct the images was considerably longer than for the prints and 



310 Pratt Institute: A Historical Snapshot of Campus and Area 

plates. The negatives, being film, showed dust and scratches much 
more prominently than the other formats, especially at the high 
resolutions we were using. Our Epson scanner had no Digital Image 
Correction and Enhancement (ICE) functionality for filtering, and we 
chose not to use the dust-removal feature native to the Epson scanner 
because we feared it was insufficiently robust. That meant depending 
on the tedious and time-consuming process of editing out dust and 
scratches by hand via Photoshop's clone stamp tool. In retrospect, it 
might have been better for us to use Epson's dust-removal feature- 
sophisticated or not— for the sake of saving much-needed time. 

The bookplates in particular presented problems regarding 
accuracy of tone and color reproduction. Following the National 
Archives and Records Administration's recommendations in 
Guidelines for Digitizing Archival Materials for Electronic Access 
(Puglia, Reed & Rhodes, 2004), we included a Kodak Q-13 gray scale 
target with each scan, which improved the accuracy and consistency of 
color and tone reproduction. We also used ICC color profiles, 
converting the ColorSync Workflow profile (standard for Epson 
scanners) to the Adobe RGB 1998 workspace for post-processing in 
Adobe Photoshop. We were not entirely satisfied with the results, and 
realized that color management software would have been useful for 
profiling and calibrating our scanners and monitors. In addition, the 
presence of several large windows in the Visual Resources Center 
meant that lighting was highly variable and rendered difficult the 
comparison of the originals against the digitized images to check 
quality and accuracy for color, brightness, and tone. The process was 
less consistent and more subjective than we would have preferred, but 
that was a drawback of being housed in a building constructed in 1896 
with windows intended to provide as much natural illumination as 
possible. 

Another obstacle was the condition of many of the items, 
especially the photographs, some of which dated from the nineteenth 
and early-twentieth centuries and suffered from cracks, brittle paper, 
peeling emulsions, and other damage. They were housed in archival 
polyethylene or Melinex sleeves for protection, which meant the 
scanning technician had to remove and replace the images from the 



Pratt Institute: A Historical Snapshot of Campus and Area 311 

sleeves carefully to avoid inflicting further damage. The technician, 
after scanning, made attempts in Photoshop to minimize some of the 
visual imperfections that appeared in the scans because of the 
condition of the originals. This was a time-consuming process. 
However, this project was as much about preservation as it was access 
and outreach. Several images were chosen precisely because they 
would not last much longer. 

One positive result of these problems was the creation of 
standards for imaging workflow and various file-management issues 
such as naming conventions and folder hierarchy systems. In one way, 
the scanning issues we had to confront were beneficial (albeit 
annoying), because they provided us with a solid foundation of 
practical experience working with a variety of formats that would 
allow us in future projects to anticipate potential problems and enable 
us to construct schedules that would accommodate them. They also 
gave us a greater understanding of the logistics involved in the 
deceptively simple and clear-cut process of "digitization," which 
increasingly has become, for the uninitiated, a buzzword for quick- 
and-dirty reproduction and document delivery. 

Metadata 

The nature of our project also contributed challenges for 
metadata creation. The diverse nature of the items required the 
cataloger to quickly gain expertise in cataloging bookplates, fashion 
plates, and historical photographs and negatives. Research performed 
in one subject area did not usually apply to another; for instance, 
knowledge of the history of Pratt Institute, necessary for the prints 
and negatives, had no bearing on the fashion plates and bookplates. 
As Pratt's archivist, I was the main source of information for questions 
about the historical images, but researching the bookplates and 
fashion plates was a bit more complicated. The cataloger had to 
consult resources at Columbia University to obtain background 
information on the bookplates, their creators, and the historical 
techniques used to create them. 

The cataloger had to use a slightly different approach for each 
type of material. The format differences across collections were a 



312 Pratt Institute: A Historical Snapshot of Campus and Area 

given, but the interpretation of the creator, description, and subject 
fields often varied as well. For example, the description field for a 
fashion plate with printed text in French required a different format 
and style than that for a photograph of an architectural drawing. In 
hindsight, it would have been more efficient and cost-effective had we 
focused on the metadata for one type of material at a time. 

As they did with color reproduction, the bookplates proved 
especially problematic. Many of the artists were identified by only a 
monogram or other symbol, and a large number were undated, but the 
cataloger simply lacked the time to research and identify each plate 
more thoroughly. The techniques used to produce the plates (such as 
etching or engraving) might also have been determined for many of 
the plates had there been additional time or cataloging resources, 
though the cataloger was able to do this for some of the plates. There 
were issues at the quality-control stage as well. For example, we 
initially entered the bookplate owners' names in the format Firstname 
Lastname. When we realized that this produced idiosyncratic sorting 
in our image presentation software, however, we changed the order to 
Lastname, Firstname— a relatively small change in the metadata 
template that proved time consuming because it had to be 
implemented in each individual record. 

We also had to modify the Dublin Core metadata template in 
order to provide more effective descriptions of the bookplates' unique 
elements. For instance, the person for whom a bookplate was made is 
as important as the artist who created it, and the motto, when there is 
one, becomes a major means of identification. We had to 
accommodate these features by creating the elements Owner and 
Motto, and we changed the Creator field to Artist because we felt 
many people might misinterpret Creator to refer to the printer of the 
physical object rather than the artist who designed the plate. 
Additional modifications were also made (Figure PRAT-2). 



Pratt Institute: A Historical Snapshot of Campus and Area 313 



Simple Dublin Core Fields 


Modified Dublin Core Fields 
for Bookplates 


Identifier 


Identifier 


Source 


Source 


Date 


Date Original 


Date Digital 


Creator 


Artist 


Title 


Owner 


Coverage 


Place 


Description 


Description 


Subject 


Subject 


Language 


Language 


Format 


Format 


Type 


Type 


Rights 


Rights 


Contributor 


Contri bating Institution 


Publisher 




Relation 






Motto 




Keywords 




Digitization Specifications 




Digitizer 




Cataloguer 



Figure PRAT- 2. Basic and Modified Dublin Core Metadata. 

Access to Our Digital Collection 

Although many of the grant recipients for METRO'S 2005 
initiative used CONTENTdm as their image management and 
presentation software, Pratt had made a decision, prior to and 
independent of the grant, to purchase Luna Insight software, which 
was used by the Getty Museum and other important art institutions. 
The reasons for this had to do with numerous organizational and 
storage features deemed necessary for the needs of our faculty and 
students: unlike some institutions, Pratt's primary and overarching 
digitization concerns had to focus on the continuing research and 
pedagogical needs of art, design, and architecture programs at both 
the undergraduate and graduate levels. This decision was not without 



314 Pratt Institute: A Historical Snapshot of Campus and Area 

consequences when it came to creating access to our METRO 
collection. 

To begin with, implementation of Luna Insight into the browser 
environment was a slow, difficult process; a significant learning curve 
was required to resolve issues surrounding Insight's interface and 
layout, which was new to everyone. Our greatest difficulty, however, 
was making the collections OAI-harvestable. Our systems librarian 
joined us at this point to provide assistance, and he discovered that 
Luna Insight's databases were not fully compliant because they were 
unable to allow selective harvesting based on a specific date stamp- 
that is, they could harvest repository files that were static but not 
dynamic. Even achieving this level of compliancy was frustrating and 
time consuming because OAI harvesting was naturally not Luna's 
responsibility (in fact, Luna's support staff told us that Pratt was its 
first client utilizing OAI harvesting for Insight collections). Therefore, 
it became our sole responsibility to generate the files, intermediate 
them with a gateway, and register them with a harvester, although no 
member of the team had done this before. 

With some difficulty we were able to generate static repository 
files with the Luna Insight software, after which we had the collections 
registered and intermediated through the Los Alamos National 
Laboratory's (LANL) open-source experimental OAI Static Repository 
Gateway software (Srepod). This was not an easy process because our 
inexperience with OAI gateway software resulted in a steep learning 
curve. Once we had gained sufficient mastery of the software, 
however, we chose to register our collections with OAIster, though 
here, too, we encountered difficulty, because our attempts were met 
with persistent error messages. Eventually we had to contact a 
member of the OAIster team at the University of Michigan, who was 
able to assure us that our collection had been registered successfully. 

What made the OAI harvesting process so challenging was the 
fact that it comprised three distinct steps and was supported by three 
independent systems: Luna Insight, Srepod, and OAIster. Luna was 
unable to help us with Srepod and OAIster, Srepod was unable to help 
us with OAIster, and OAIster was unable to help us with Srepod, yet 



Pratt Institute: A Historical Snapshot of Campus and Area 315 



the three steps of the process had to mesh with one another in order to 
achieve success. Moreover, as open-source systems, Srepod and 
OAIster understandably could not provide extensive personalized 
technical support. 

We also encountered compatibility issues because, as mentioned 
above, we modified the Dublin Core template when cataloging the 
bookplates. During the harvesting process, we discovered that OAI 
gateways and harvesters require strict compliance of metadata with 
the simple Dublin Core schema (i.e., the basic fifteen fields). We were 
able to solve this problem by retaining our specialized, bookplate- 
specific fields while adding the standard fields we had previously 
excluded. We thus ended up with two metadata templates— the 
specialized fields and the standard Dublin Core fields— in the same 
record, with duplicated values in some of the fields (for example, the 
values in our specialized Artist field were duplicated exactly in the 
Dublin Core Creator field). In our Luna Insight presentation, only the 
specialized fields were displayed; when we harvested for OAI, only the 
Dublin Core fields were harvested. 

Perhaps no other problems encountered during this project were 
as perplexing and complicated as those surrounding OAI compliancy. 
Confronting these issues provided us with a greater understanding 
and appreciation of the complexity of the process, but it also brought 
home the fact that, without considerable personalized technical 
support, it was difficult to fulfill an important requirement of the 
grant initiative. We also understood that it would be unfair to expect 
any system to be able to provide that level of guidance. It was very 
much a learning experience, and our systems librarian documented 
each step of the process, from the generation of the static repository 
files with the Luna Insight software to the final registration of the 
collection with OAIster. 

2006 and Beyond 

As we approached the completion of the 2005 project, our director 
urged us to apply for another grant in 2006 to maintain our 
momentum. This time, we focused on only one theme. One advantage 



316 Pratt Institute: A Historical Snapshot of Campus and Area 

of our previous metadata research was the discovery that our 
bookplate collection included work by several important American 
bookplate artists, and we also recognized the general research value of 
bookplates for librarians, antiquarian booksellers, and genealogists, as 
well as their beauty, creativity, and technical achievements. Therefore, 
we decided to digitize the remaining 1,100 plates in our ex libris 
collection, anticipating that the digital dissemination of the plates 
would serve not only the Pratt community but also the increasing 
number of individuals and organizations, both national and 
international, dedicated to the study of ex libris. When METRO 
awarded us our second grant, we were delighted with the opportunity 
to digitize an entire collection— and one, moreover, that had lain 
hidden in Special Collections since the late 1970s. For almost three 
decades no one had known about or been able to access these 
bookplates; now, anyone with Internet access would be able to. 

Although several individuals on the 2005 team had left, a few 
remained and, with me as project manager, formed the nucleus of the 
2006 group. Despite the personnel changes, however, this project 
proceeded more smoothly than the previous one, partly because we 
applied the lessons learned from the 2005 project and partly because 
we were dealing with only one collection (for which we had already 
worked out metadata issues) and therefore did not have to cope with 
multiple formats and approaches. Despite our allowing more time to 
import the images and metadata, we once again found it necessary to 
readjust our timeline because it was not easy to calculate accurately 
the amount of time necessary to digitize, enhance, and catalog the 
plates, even though we knew in advance how many there were. Also, 
we initially thought we could jump start the cataloging by initiating 
metadata creation before scanning, but we soon were forced to reverse 
the order, because the cataloger needed to zoom in on many images to 
determine certain details for the metadata (for instance, whether the 
plates were etchings or engravings). In addition, some of the scanning 
problems described for 2005 had not been completely corrected at the 
time we began the 2006 project, and so it required additional time to 
resolve them finally. Fortunately, thanks to the cushion we had built 



Pratt Institute: A Historical Snapshot of Campus and Area 317 

into our original timeline, the schedule modifications did not affect 
the completion of the project (Figure PRAT- 3). 

Once our online collections went live, we publicized them 
throughout the Institute and held an open house to provide 
instruction and hands-on demonstrations. 



Promotion 


Content 
Import 


s 


Scanning/ 
Imaging 


Preparation 




REVISED DIGITIZATION PROJECT TIMELINE 












90/10 
























03/06 












04/06 












05/06 












S> 












-=■ 

O 












CD 

c5 

ON 
























90/01 












err 
T> 


J 













Promotion 


0 s= 
S. S- 


Scanning/ 
Imaging 


Metadata 


Preparation 




ORIGINAL DIGITIZATION PROJECT TIMELINE 












3N 












= 

KJ 

3N 












i5 

3N 












04/06 












Si 

IN 













CD 

=N 












90/i0 












g 












09/06 












10/06 



























Figure PRAT-3. Schedule for the 2006 Grant. 



318 Pratt Institute: A Historical Snapshot of Campus and Area 

We also notified various American and international 
organizations dedicated to the study of bookplates, such as the 
International Federation of Ex Libris Societies (FISAE), because we 
anticipated (correctly) that our collection would attract considerable 
long-distance interest, and hoped that scholars more knowledgeable 
than we would contact us with additions and corrections to our 
metadata. Several sites posted our announcement, and we were 
especially pleased to hear from the Australian Bookplate Society, 
which noticed that Australia was not represented in our collection and 
actually offered us a donation to fill that gap! 

In 2008 the Pratt Institute Libraries chose to become a 
participating institution in ARTstor, which meant, for us, 
discontinuing Luna Insight. Our art, design, and architecture images 
would be hosted as a local collection through ARTstor, but that was 
not an appropriate repository for those from Special Collections and 
the archives because of their unique and (in the case of the 
photographs and negatives) proprietary nature. The limitations on our 
financial resources meant that whatever image management software 
we selected would have to be open source, but time constraints 
prevented us from thoroughly examining the variety of available 
systems, so our Visual Resources Curator suggested we post them on 
Flickr. This turned out to be an excellent idea, because we very quickly 
saw a marked increase in visitors, some of whom supplied information 
that, for our bookplates, allowed us to expand or correct our 
cataloging, which was precisely what we had been hoping for. The 
collection logged over 300,000 views in the first nine months on 
Flickr, and not a single image has been viewed less than nineteen 
times! Clearly our presence on Flickr allowed individuals to discover 
us serendipitously in a way not possible when they were required to 
navigate our Website. 

I noticed, too, that staff from the Institute's administrative 
branches, such as the Public Relations and Communications Office 
and the Office of Alumni Relations and Annual Giving, more 
frequently browsed the collections before making requests for 
historical images, with the result that they knew exactly what they 
wanted and were even able to provide us with source and identifier 



Pratt Institute: A Historical Snapshot of Campus and Area 319 



numbers. This has increased the efficiency of our workflow and 
document-delivery capability enormously. 

The only drawback with Flickr is its inability to support OAI- 
compliancy. We are hoping to acquire an open-source image 
management system that will correct this. Meanwhile, we continue to 
digitize archival images as needed, and have even begun digitizing 
important historical publications frequently used by researchers, 
which we are planning to post on the Archives page of the Libraries' 
Website. There is no question but that we have experienced a sea 
change in our digitization efforts as a result of our two grants, and our 
story vividly illustrates not only the old maxim that success breeds 
success, but also that long-term impetus can be generated from a 
single opportunity. 

Acknowledgments: The success of our projects was achieved only 
because of the dedication and expertise of the members of the project 
teams, whose members included Amanda Schriber, Brian Cross, 
Lorraine Smith, Vernon Bigman, Stephen Klein, and Michael Nolasco. 
I am particularly grateful to Lorraine Smith, until recently the Visual 
Resources Curator at the Pratt Institute Libraries, who was our image 
cataloger for the 2005 project and supervised the digitization process 
in 2006 as Acting Visual Resources Curator. She was responsible for 
coping with many of the technical issues described above and 
reviewed this chapter to make sure my explanations were accurate. 

References 

Bibliographic Center for Research [BCR]. (2008). BCR's CDP digital 
imaging best practices, version 2.0. Retrieved from 
http://bcr.org/dps/cdp/best/digital-imaging-bp.pdf 

Gertz, J. (2007). Preservation and Selection for Digitization 

[Northeast Document Conservation Center Preservation Leaflet 
No. 6.6]. Retrieved from 

http://www.nedcc.org/resources/leaflets/6Reformatting/06Preser 
vationAndSelection.php 



320 Pratt Institute: A Historical Snapshot of Campus and Area 

Hagen, D., Horrell, J. & Merrill-Oldham, J. (1998). Selecting research 
collections for digitization. Retrieved from 
http : / / www. clir.org/ pubs/ reports /hagen/ pub74.html 

NC ECHO. (2007). NCECHO guidelines for digitization 2007 (Rev. 
ed.). Retrieved from http://www.ncecho.org/dig/ 
digguidelines . shtml 

Puglia, S., Reed, J. & Rhodes, E. (2004). NARA technical guidelines 
for digitizing archival materials for electronic access: Creation of 
production master files— raster images. Retrieved from 
http://www.archives.gov/preservation/technical/guidelines.pdf 

Southeastern NY Library Resources Council [SENYLRC]. (2004). 
General collection criteria guidelines for HRVH. Retrieved from 
http:/ /www.hrvh.org/ about/ selectioncriteria.htm 



Scaling Back for an "Experimental" 
Collection 



Mark F. Anderson (The University of Iowa Libraries) 



Abstract 

Digital Library Services (DLS) at the University of Iowa Libraries has 
progressively worked toward coordinating more large-scale 
digitization projects both within the libraries and across campus, 
moving away from model of web exhibits that were often created 
before the department was formed in 2005. However, a variety of 
situations still call for small-scale projects. This chapter, describing 
the design and production of the "W9XK Experimental Television 
Digital Collection", shows that small-scale digitization projects can 
bridge that gap, and yield collections that rise above the level of web 
exhibits in their usefulness to scholars and the general public by 
limiting exclusive selection and promoting comprehensiveness. While 
mirroring this approach of mass-digitization, digital librarians can 
also use curatorial decisions and software functionality to further 
assist users of these small-scale collections. 

Keywords: Archives, Digital collections, Digital libraries, 
Digitization, Web exhibits. 



Introduction 

Henry Wadsworth Longfellow wrote, "Most people would succeed in 
small things if they were not troubled with great ambitions." Even if 
the poet was not referring specifically to digital initiatives in libraries, 
it's an appropriate sentiment at a time when so much institutional 



321 



322 Scaling Back for an "Experimental" Collection 

effort is directed toward mass digitization projects such as the 
agreement between Google and the Committee on Institutional 
Cooperation (the consortium to which The University of Iowa 
belongs) to digitize no less than 10 million volumes from among 
members' collections (Committee on Institutional Cooperation, n.d.). 
At the same time, it has become common for institutions, or units 
within institutions, to work on building small, narrowly-focused 
collections as an initial foray into digitization, while securing funding 
for the necessary resources to ramp up to large-scale scanning 
projects and mass digitization, or to expose exceptional materials. 
This chapter will discuss one of these small digital collections, but will 
begin with the development of the digital library at the University of 
Iowa to which it belongs. 

Digital Initiatives at the University of Iowa 

The earliest digital collections hosted by The University of Iowa 
Libraries date back to the mid-1990s, and were a combination of small 
web exhibits and larger efforts to digitize materials from flagship 
collections. The web exhibits were considered "virtual versions" of 
exhibits displayed in the library's exhibit hall and which combined 
highlight images with essays and other supporting text provided by 
library staff and campus scholars (University Libraries. University of 
Iowa, 1999). In 1997, the Libraries were awarded Library of 
Congress/Ameritech National Digital Library Competition funds, 
which supported the digitization of nearly 8,000 talent brochures 
from the papers of the Redpath Lyceum Bureau, an agency that 
represented performers on the Midwest Chautauqua circuit. In many 
ways, this was the Libraries' first mass-digitization project (The 
Library of Congress, 2003). Other scholarly digital collections created 
at the time included the Center for Electronic Resources in African 
Studies and the International Dada Archive. 

In 2005, a new department, Digital Library Services (DLS), was 
formed to manage digital initiatives. Since that time, DLS has 
progressively worked toward coordinating more large-scale, "left-to- 
right" projects both inside the Libraries as well as with faculty and 



Scaling Back for an "Experimental" Collection 323 



scholars on campus, moving away from the highly selective model of 
the early web exhibits, which tended to contextualize the digital 
materials for the user. Conversely, the model of comprehensive 
digitization aims to allow users the freedom to repurpose digitized 
materials for the creation of new knowledge and insight. This requires 
not only a thorough approach to digitization on the part of the library, 
but also the availability of tools to facilitate new uses for the items, 
especially tools that link, overlay, and share data. Unfortunately, many 
of these tools require extensive development and are not yet available 
as turn-key solutions. 

Situations such as the anniversary of an important institutional 
event or a specific user-driven request have still led libraries to 
undertake small- and medium-sized digitization projects. The 
challenge for libraries is to create smaller digital collections that are 
valuable for scholarship. Small-scale digital collections may provide 
the same research potential as the entire physical collections from 
which they are drawn. 

Within the University of Iowa Libraries, DLS has cultivated a 
particularly close relationship with three content providers: Special 
Collections, Iowa Women's Archives, and University Archives. With 
the first two units, digitization projects are becoming more large-scale 
and left-to-right in their approach, but less so with the latter. This is 
mostly due to that fact that University Archives subscribes to the 
"principle of provenance", the archival practice of organizing 
collections by creator rather than subject, and collections are often 
selected for digitization based on a topic of interest rather than 
contributing entity, making it difficult to consider left-to-right 
scanning. For physical exhibits of University Archives materials, this 
has long led to "artificial collections" that pull from many different 
record groups, according to University of Iowa Archivist David 
McCartney. More recent digital collections from the archives are no 
exception. (D. McCartney, personal communication, October 27, 
2009). This kind of intermediation by the selector is a major 
difference between the small-scale "boutique" digitization model and 
mass-digitization. 



324 Scaling Back for an "Experimental" Collection 

In "Shifting Gears: Gearing Up to Get Into the Flow," the authors 
direct libraries to stop obsessing about items and making curatorial 
decisions about what to digitize since the selection has already been 
done (Erway, R., and Schaffner J., 2007). Arguing that if items were 
important enough to acquire originally, they are important enough to 
digitize, the authors advocate a more programmatic approach of 
digitizing items as they are accessioned, rather than thinking of 
digitization in terms of discrete projects. Is it possible to work within 
these provocative guidelines and still develop small-scale digital 
collections that are "in the flow"? Yes. For example, libraries can push 
the decision making process from the item level to the collection level. 
After all, with so much catching up to do, there's no way all existing 
collections can all be scanned at once. Small collections can be 
selected for digitization, and worked through left-to-right, even if that 
means scanning hand-written grocery lists on backs of dry-cleaning 
receipts (a real-world example from an Iowa Digital Library 
collection). In this way, DLS has operated with the understanding that 
a digital collection, even a small one, should be a comprehensive 
surrogate of a physical collection, regardless of size or scope, which 
attracts not only general interest, but promotes new scholarly 
activities as well. Also, for text items, the combination of brief 
metadata and optical character recognition (OCR) can provide access 
without the need of extensive, handcrafted records. 

Selecting the Collection 

At a time when video is so easily accessible through websites Hulu, 
Netflix, Boxee.tv, and YouTube as well as thousands of cable and 
satellite television stations, it's hard to imagine the technical and 
organizational challenges facing engineers and technologists involved 
in the infancy of television. Were these pioneers of a century ago 
similarly considering the scale of their new experimental medium? If 
so, then the small digital collection described in this chapter is 
particularly appropriate: "W9XK Experimental Television at Iowa" 
(http:/ / digital.lib.ui0wa.edu/w9xk). 



Scaling Back for an "Experimental" Collection 325 



February 2009 was seen as an excellent opportunity to promote 
the history of the W9XK experimental television station on the 
University of Iowa campus by means of digitizing related archival 
materials, both because of the congressionally-mandated conversion 
to digital broadcast television, which was set to take place that month, 
and the 75th anniversary of the station the year before. W9XK, the 
first educational TV station in the U.S. went on the air January 25, 
1933> with a weekly or twice-weekly schedule of lectures, music, and 
drama. Reception was reported back to the university from as far away 
as Oklahoma, Kentucky and New York (University Libraries. 
University of Iowa, 2009). Appropriate materials were gathered from 
throughout the archives including correspondence and newspaper 
clippings from faculty and electrical engineering department subject 
vertical files, still photographs from a prominent campus 
photographer, articles from a campus engineering journal, WPA 
federal writers project papers, and a student thesis. 

With no single collection to process in a left-to-right manner, this 
project is a departure from DLS's more recent production model 
described earlier, and runs somewhat contrary to the challenges in 
"Shifting Up," but it is unlikely that these materials would be scanned 
together in any other way short of a heroic effort to mass-digitize the 
entire University Archives. The collection is small, containing just 
over 50 digital objects (although some of the objects are composed of 
several items grouped together, their organization to be discussed 
below), but it tells an important story of the station's development and 
operation in the broader context of early advances in television 
technology, using the variety of formats listed above. Unfortunately, 
no video or film footage of \V9XICs broadcasts could be included in 
the collection since none is known to exist today. 

Building the Collection 

From 2005 to 2007, digital initiatives at the University of Iowa 
Libraries were completed mostly by staff in DLS in "in-house" 
production. Scaling up to support larger digitization efforts has 
required production tasks to occur in other library departments with 



326 Scaling Back for an "Experimental" Collection 

both the staffing and expertise to process not only more materials, but 
also address unique handling and cataloging issues that invariably 
arise from digitizing entire collections. Metadata is now mostly 
applied by cataloging staff, except in cases where collections come to 
the library already packaged with existing records, which requires 
more in the way of reorganization and standardization rather than 
creating new records from scratch, which as a brand new collection, is 
the way metadata for the W9XK collection was handled. 

In a similar way, the scanning that had been done by staff and 
students working directly under DLS has mostly moved to the 
Preservation and Reformatting department. However, scanning for 
the W9XK collection broke from this workflow model. A staff member 
in University Archives was interested in learning about scanning and 
image editing, especially the operation of a new top-down book 
scanner the Libraries had just purchased. The potentially small size of 
the digital collection presented a good opportunity for that. In this 
way, small-scale digital collections afford opportunities to deviate 
from established workflows for staff to develop new skills. But with 
other parts of production spread out, it may not make a considerable 
difference in time to do just part of the production "in-house". The 
time for an item in the W9XK collection to be completely processed 
and added to the digital collection was neither considerably faster nor 
slower than other collections. 

The University Archivist selected and pulled appropriate 
materials from the physical collections listed above and prepared 
them for the staff member who scanned them on either an 11x17 
flatbed scanner or the newly acquired top-down scanner, which 
accommodates items up to 24 x 36 with a book cradle for bound 
materials. Scans were then edited and display files derived by DLS 
staff who then uploaded the images to CONTENTdm, the digital 
library software that the Libraries use for many of its digital 
collections. Lastly, cataloging staff added descriptive, administrative 
and technical metadata to the objects. Metadata is almost always 
applied to the digital items following upload to the digital content 
management system in order to take advantage of software 



Scaling Back for an "Experimental" Collection 327 



functionality that allows for easier batch cataloging processes, 
increasing efficiency. 

Collection highlights and sample searches were selected for a 
collection homepage, and the university archivist provided an 
introductory paragraph, which takes the place of the more contextual 
and interpretative writing that used to accompany web exhibits 
directly alongside images. A lingering concern about small-scale 
digital collections is packaging them in beautiful, but deep, web 
portals that users must navigate to find the information they seek, 
rather than letting digital objects live in a large repository with thin 
logical collections built around them. A compromise is to maximize 
collection exposure to search engines and aggregators. The Libraries 
use Google Sitemaps to help assure collections are crawled and ranked 
highly by Google. OAI harvesting of collections is enabled for outside 
service providers, and the Libraries also implemented a next- 
generation catalog that brings together records from the traditional 
catalog as well as online books and journals, local web pages and 
digital collections such as W9XK. 

One recent change in how DLS delivers large-scale, left-to-right 
digital collections is to use folder-level organization of the digital 
objects. The digital library software used for most collections in the 
Iowa Digital Library, CONTENTdm, uses "compound objects" to 
deliver multi-page items, from the two sides of a postcard, pages of a 
document, a monograph, or an entire folder's worth of materials 
arranged hierarchically. This decreases the number of items for users 
to browse through up front, or consider in results sets. While the 
W9XK collection did not fit the folder-structure organization model 
since its items came from many different archival collections, one of 
the digital collection-delivery decisions made was to display 
correspondence items (letters between engineering faculty, university 
administrators, government authorities, and the public) so that letters 
and their replies were in the same compound objects, rather than as 
separate objects. One could argue that this mediation is too much 
along the lines of a web exhibit, but it was thought to be helpful in 
navigating the correspondence. Feedback from users would help to 
support whether this is a good decision. In any case, digital librarians 



328 Scaling Back for an "Experimental" Collection 

often have to make organizational decisions based on software 
functionality and with the intention of assisting users and researchers. 

Use of the Collection 

The W9XK Digital Collection was officially launched on February 17, 
2009 through press releases and library blogs. While we must expect 
that our users will tend to find our digital collections through the back 
doors of search engines and aggregators rather than our finely-crafted 
web portals, only 10% of visits have come directly through Google, 
while more than half of all users are referred to the collection from 
promotion on university and library pages as well as the library 
catalog, although the search engine could have sent users to these 
local pages first. Over 75% of page views to date occurred during the 
first two months, owing much to the initial publicity (Figure IOWA-i). 

3000 
2500 
2000 
1500 
1000 
500 
0 

Feb. Mar Apr May June July Aug Sept Oct Nov 



Figure IOWA-1. W9XK Pageviews by month (2009) 

While somewhat discouraging, most early use may have just been 
casual interest, and it may take longer for research and scholarly use 
to increase. Owing to Lally and Dunfords assessment of the impact of 
Wikipedia in directing users to library digital collections, links to the 
W9XK collection have recently been added to Wikipedia articles on 
experimental television and may have an effect on traffic over time. In 
"Using Wikipedia to Extend Digital Collections" (Lally, A. and 
Dunford, C, 2007), the authors note that Wikipedia is "a prime 




Scaling Back for an "Experimental" Collection 329 



candidate for ... efforts at pushing information about the Libraries out 
to where users conduct their research," as their server statistics 
indicate it is indeed driving a significant amount of traffic to their 
collections. 

65% of visits to the W9XK collection home page have come from 
within the state of Iowa, a higher percentage of local visits than for 
two other small-scale digital collections in the Iowa Digital Library, 
which were prompted by high use and requests for reproduction from 
unique physical collections, especially where appeal went beyond the 
local area (Figure IOWA-2). 



■ W9XK □ Victor Projectors □ Rand & Leopold Desks 





















27 








9 









Figure IOWA-2. Percent collection use from Iowa networks 

Two examples of these are the Victor Animatograph pamphlets 
(http://digital.lib. uiowa.edu/victor) and the Rand and Leopold Desk 
catalogs ((http:// digital.lib.uiowa.edu/desks), both with fewer than 
30 items each. The Victor Animatograph Corporation of Davenport, 
Iowa made the world's first 16mm motion picture projector, and was a 
leader in the area of film technology throughout the 20th century. 
People in possession of one of these projectors would often contact the 
Libraries requesting copies of the product pamphlets, which contain 
detailed schematic diagrams and operating instructions. This 
collection was built before resources allowed for left-to-right 
digitization, so rather than digitize the entire collection of papers from 
the Victor Corporation, just the 21 highly-requested pamphlets were 
scanned. Since launching the collection, 20% of visits to the collection 
have come from outside the United States, and over 70% of visits to 



330 Scaling Back for an "Experimental" Collection 



the collection have come from outside the state of Iowa according to 
site statistics. 

The Rand & Leopold Desk Company was an office furniture 
manufacturing firm which operated out of Burlington, Iowa for 102 
years before closing in 1990. Furniture collectors had always been 
interested in the Libraries' collection of 27 original product catalogs, 
and there were many requests for reproduction. After scanning them 
and making them available through the Iowa Digital Library, only 7% 
of site visits to the digital collection have come from outside the US, 
but over 90% have come from outside the state of Iowa. Perhaps more 
so than the W9XK collection, these two small collections meet the goal 
of facilitating information needs of users who cannot easily travel to 
the UI Libraries. 

Difficulties in Tying a Bow 

Even when original goals are met, small-scale digital collections don't 
often stay "done" for long. Scope-creep is a common danger, for even 
when items haven't been cherry-picked for a digital collection, the 
source collection almost always logically connects with more 
collections. In trying to connect users with a broader universe of 
resources, libraries are compelled to extend the scope of their digital 
collections. In an example of how digital collections never seem to stay 
finished, a collection of an artist's papers in the Iowa Women's 
Archives were digitized to complement her digitized artwork. When 
discussing publicity of the collection, IWA staff noted that they had 
found several audio oral histories and also a video made about the 
artist. Even when those multimedia materials are digitized and added 
to the collection, it's possible that the archive will acquire additional 
items that would need to be added to the digital collection in order for 
it to remain comprehensive. 

Conclusions 

Even as mass-digitization projects begin and are sustained at libraries 
through funding and organizational commitments to scaled-up 
scanning workflows, requests for creating small-scale digital 



Scaling Back for an "Experimental" Collection 331 



collections will persist. This is because many unique collections, 
especially those with preservation issues or spread out in archives will 
not be included in mass-digitization workflows. Likewise, the needs of 
libraries' clientele and donor relationships do not necessarily coincide 
with broad institutional objectives. Responding to these requests can 
help strengthen relationships inside and outside of the library and 
build support for digitization programs. Even one-off collections 
should fit within programmatic guidelines, although in the case of the 
W9XK collection, some liberties were taken in its creation. Small-scale 
digital collections can be useful to researchers and the general public, 
but they have to be accessible. Once found, the collections should be 
complete to the point where users can be confident that they have the 
entirety of a collection's information at their disposal. Although the 
W9XK digital collection required the hand-selection of items, steps 
were taken to ensure the fullness that is necessary in today's digital 
libraries. 

References 

Committee on Institutional Cooperation, (n.d.). CIC-Google 

agreement. Retrieved from http://www.cic.net/Home/Projects/ 
Library/BookSearch/CIC-Google.aspx 

Erway, R., and Schaffner, J. (2007). Shifting gears: Gearing up to get 
into the flow. Retrieved from http://www.oclc.org/ programs/ 
publications/reports/2007-02.pdf 

Lally, A., and Dunford,C. (2007). Using wikipedia to extend digital 
collections. D-Lib Magazine, 13 (5/6). 

Library of Congress. (2003). Library of Congress and Ameritech 
Competition (igg6-iggg): Awards and collections. Retrieved from 
http:/ /lcweb4.1oc.gov/ ammem/ award/ collections.html 

University Libraries. University of Iowa (1999). Keeping our word / 
North Lobby exhibit / University of Iowa. Retrieved from 
http : / / www.lib.uiowa.edu / exhibits /keeping/ intro 1 .htm 

University Libraries. University of Iowa. (2009). Iowa city town and 
campus scenes. Retrieved from http:/ / digital.lib.uiowa.edu/ictcs 



332 Scaling Back for an "Experimental" Collection 

University Libraries. University of Iowa. (2009). Victor Animatograph 
Company Pamphlets. Retrieved from http://digital.lib.uiowa.edu 
/victor 

University Libraries. University of Iowa. (2009). Map/diagram of 
Midwest, 1930s. Retrieved from http://digital.lib.uiowa.edu/ 
u?/w9xk,67 

University Libraries. University of Iowa. (2009). Rand and Leopold 
Desk Company Catalogs. Retrieved from 
http:/ / digital.lib.uiowa.edu/ desks 



From Confusion and Chaos to Clarity and 
Hope: Reorganization of Work Flows, 
Processes, and Delivery for Digital 
Libraries 

Jody L. DeRidder (The University of Alabama Libraries) 



Abstract 

Digitization support within an institution may be fractured across 
several departments, only partially funded, and may suffer restraints 
imposed by delivery software which seriously hamper progress. Most 
digitization is undertaken with little thought for the future; the result 
is digital file chaos and confusion. Without clarification of file 
identities and relationships, preservation and migration to new 
systems are seriously hampered. Additionally, low funding for archival 
staff may preclude the creation of valuable item-level metadata. The 
University of Alabama Libraries leveraged the expertise available 
across the library to build a cross-departmental collaboration with 
which to face our challenges, recognizing that obstacles become 
opportunities for creative solutions. We are involved in a series of 
pilot projects to explore how to address the gap in archivist staffing to 
create item-level metadata. This chapter shares our discoveries and 
solutions. 

Keywords: Cross-departmental collaboration, Digitization, Digital 
file organization, Metadata creation, Open source software. 

The first few years of most digital library initiatives are marked by 
'boutique' collection development, in which the standards, 



333 



334 From Confusion and Chaos to Clarity and Hope 

organization, methodology, metadata, file names, and consistency 
vary considerably. At the time of my arrival at the University of 
Alabama in mid-2008 as head of the new Digital Services department, 
over thirty such digitization projects had been completed. Each 
collection had its own file-naming system and metadata fields, with 
inconsistencies throughout; nothing was standardized. Metadata in 
the delivery software did not retain in any predictable fashion a 
reference to the related archival files, and could not be exported in 
full. Digital Services staffing was minimal, requiring time from the 
Cataloging and Metadata Services for subject headings and upload, 
from archival staff for preparation of content and descriptions, and 
from Web Services to manage the interface and software support. 

The scope of the task ahead was to expand heavily on the 
scanning staff and equipment, develop a feasible set of systematic 
work flows for supporting a large increase in scanning, build a cross- 
departmental team capable of supporting digital library development, 
and to create an organized and reusable set of digital content that is 
not dependent upon resident knowledge for continuation or 
restoration. Challenges included a simultaneous reduction in archivist 
work hours, minimal space for expansion, difficult relations between 
some departments, and insufficient time available from Web Services. 

As in many smaller organizations, our digitization effort is 
tremendously dependent upon cross-departmental collaboration. 
Programming assistance and web delivery, metadata services, 
archivist expertise and a regular influx of well-chosen content are all 
critical to the development of our online research collections. A 
previous gift from EBSCO Industries to the libraries supports 
digitization and the development of technical infrastructure, but it 
does not support the processing, arranging, and description of 
archival collections. Our need for content creates a demand on the 
archivists that they simply do not have the resources to meet. 

Recognizing the need for improved cross-departmental 
communications and teamwork, our dean (L. A. Pitschmann, personal 
communication, August 25, 2008) called together lead representatives 
(including two associate deans) from Library Technology, Web 



From Confusion and Chaos to Clarity and Hope 335 



Services, Collection Development, Cataloging and Metadata Services, 
Special Collections and Archives, and Digital Services, to form an 
ongoing Digital Programs group which would meet regularly to hash 
out problems, develop alternatives, research opportunities and assign 
priorities. The creation of this group was a stroke of brilliance. By 
forming this framework for participation, setting forth a strategic goal 
and providing clear administrative support, our dean laid the 
groundwork for success. Given our multiple operational and relational 
challenges, we could only succeed by seeking solutions with the 
assistance of all impacted parties. 

Against this backdrop we are working through four major 
problems: digital file chaos, the inability to reunite metadata with the 
archival content, software restrictions on the number of collections, 
and a lack of archivist time to create item-level metadata. 

Problems and Solutions 

Digital File Chaos 

Managed, efficient production and expanded growth of 
collections requires standardization, not only of work flows and 
procedures, but also of storage and file naming conventions. Delivery 
systems become outmoded; migration into replacement software 
requires consistency of legacy content. Consistency of file names and 
storage patterns can also support cost-saving automation. 

As mentioned, we already had over thirty digitized collections, 
each with different file naming systems and metadata in various states 
of disarray across a completely disorganized file system. What little 
documentation existed was scattered. Collections had been digitized 
for years with no road map, and with no concern beyond getting the 
content onto the web. We needed a clear methodology for file 
organization. 

I consulted with the archivists to gain a greater awareness of the 
scope of current and future digital content. After much debate, we 
determined that it was most important for us to store content in such 
a way that we could retrace the material to the archival analog 



336 From Confusion and Chaos to Clarity and Hope 



collection. Digital collections are transitory and overlapping by their 
very nature; we decided that the perception of a digital collection must 
be determined by metadata, not by origin (searching on a shared value 
can retrieve all components of a digital collection). Together we 
developed a file naming scheme to encompass all our holdings and 
projected digitization plans for the next few years. We created 
hierarchical levels of organization: first by holder, then by collection 
within those holdings, then by item within that collection, and finally 
by sequence for delivery (Figure ALAB-i). Each hierarchical level is 
concatenated with an underscore in the file name, so that provenance 
and location, as well as sequence for delivery, are automatable and 
clear. 



| HolderlD: u0003 
["""J Collection !D: 0000023 
item ID: 0000007 
n Sequence ID: 0005 



Archival File: u0003_0000023_0000007_00005. tif 

Figure ALAB-1 . [University of Alabama Libraries Digital File Naming Scheme. (©2009, 
University of Alabama Libraries. Used with permission.)] 

A "holder" is construed loosely, as we could not incorporate all 
the variant hierarchies of our organization feasibly into a file name. 
Supporting metadata clarifies identity and relationships. It was 
important for us to prefix each identifier with a letter so it can be used 
as an ID attribute within XML files (W3C, 2005). The n groupings are 
non-profit agencies, the p is for patron holdings, and u is for content 
from holdings areas within our university. We grouped content by 
format, as each format requires different handling and metadata 
description. For example, uoooi is the still image collections from the 



From Confusion and Chaos to Clarity and Hope 337 

Hoole Special Collections; U0002 is rare books, 110003 is manuscript 
collections, and so forth. Any of these holdings areas may have 
multiple collections. 

For the collection number, which composes the second segment 
of an identifier (concatenated with an underscore separator), we 
echoed the existing collection numbering system whenever possible. 
Thus, MS 1980 will be the collection number ending in 1980, for 
example. 

As an example, uooo3_oooo252 is the identifier for the digitized 
manuscript content of the collection known to our archivists as MS 
252. Items and their subsidiary pages, however, are numbered 
sequentially. Item numbers follow the collection number, again, 
concatenated after an underscore. For example, the fifth digitized 
letter in the MS 252 collection would be identified by 
uooo3_oooo252_oooooo5. If it's a multi-page item, there will be a 
fourth set of numbers here, one per page, to identify the sequence for 
delivery. The image for page 4 of letter 5 would be identified by 

U0003_0000252_0000005_0004. 

Thus, every part of every object has a defining and unique 
identifier which relates it to every other file in some fashion. We 
struggled with all the different anomalies we currently had and could 
foresee, simplifying this identification system as much as possible 
while still retaining the generic flexibility to apply it to all of our 
digitized content, regardless of the form or treatment. We then 
expended a great deal of effort to bring all legacy content into the new 
file organization scheme, gathering what information was still 
available to make sense of the chaos of files that remained from 
previous projects. Because of this system of file naming and 
organization, we are now able to automate much of our quality 
control. We have scripts that locate missing sequences, incorrectly 
named files, or files in the wrong place. We even have scripts to repair 
file names when large quantities of file names are in error. This has 
already saved us hundreds of hours. 

Given that each level of the organization scheme contains or 
relates information which applies to each level below it, our associate 



338 From Confusion and Chaos to Clarity and Hope 

dean (Thomas C. Wilson) decided it made sense to echo this file 
naming scheme in the directory organization for storing files. 
Everything from a particular holder should be together in a directory 
named for that holder. Everything from a particular collection should 
be in a directory named for that collection. Thus, in our storage 
directories, all of the Hoole manuscript collections will be within the 
same directory (UOO03). This is an intuitive use of the file directories 
to provide clarity and simplicity of organization. 

For example, within the UOO03 directory is a subdirectory 
0000159, which contains information about all digitized content from 
MS 159. Within this collection directory exists a subdirectory for each 
item, named for the item number. The storage directories thus echo 
the file naming scheme, providing clear, simple, automatable 
organization. Drilling down through the file system to the logical 
depth locates the digitized archival file (Figure ALBA-2). 

Metadata and documentation are stored at the levels to which 
they apply. Thus, metadata about the collection is in the Metadata 
folder at the collection level, metadata about an item at the item level, 
and metadata about a specific page is at the page level. Each sub-file 
inherits the information available at the levels above it. Thus, 
provenance documentation added at the collection or holder level 
clearly applies to all files in the directories below it. If some 
information only applies to page 4 of a letter, it is stored in that file's 
directory. 

An organizational patterning such as this (Figure AIAB-3) retains 
the item structure, both physically and nominally, through the file 
identifiers. The simplicity, systematic numbering, sequencing, and 
clear documentation stored at the level applicable makes the digital 
content resurrectable for future delivery systems, without complex 
metadata schemes or database dependence. The organizational 
scheme is built to be scalable and extensible enough to manage digital 
content into the foreseeable future. In addition, because the directory 
structure echoes the file names, we were able to automate the storage 
of content and the creation of attendant LOCKSS (LOCKSS, 2008) 
manifests. 



From Confusion and Chaos to Clarity and Hope 339 




Figure ALAB-2. University of Alabama Libraries LOCKSS Content Organization. (© 
2009, University of Alabama Libraries. Used with permission.) 



340 From Confusion and Chaos to Clarity and Hope 



V 0 U0003 

T 0 0000001 

0 0000001 
^ 0 0000002 
T 0 0000003 
^ 0 0001 
▼ 0 0002 
T 0 001 

^ 0 Metadata 

i u0003_0000001_OC00003_0002_001.tif 

▼ 0 ooz 

» 0 Metadata 

1 u0O03_Q0Q0OOl_Q0OQOQ3_O0O2_0Q2.tif 

Figure ALAB-3. University of Alabama Libraries Digital File Naming Scheme: Sub- 
page numbering. (© 2009, University of Alabama Libraries. Used with permission.) 

The manifest pages (Figure ALAB-4) link all files to be harvested 
for duplication across the Alabama Digital Preservation Network 
(Network of Alabama Academic Libraries, 2009), which is to date the 
lowest-cost model of digital preservation. 

To secure our content further, we capture MD5 checksums upon 
deposit in our storage system, and verify them weekly prior to full- 
tape backups. This enables us to ensure that the original file is 
unaltered over time, and can be restored from a good backup should 
any corruption occur. 

Reuniting Metadata with Archival Content 

Software systems come and go, and successful transition between 
them is dependent upon standardized and coherent content and 
metadata. Most of our metadata had been altered or added after 
upload of content into our delivery system, and did not exist with the 
archival files. Upon examination of the exports from our software 
delivery system, we found that no single export contained all the 
metadata for a given archival object. Even in the most complete export 
option, there was no page-level metadata apart from the title and 
transcript, and archival file references were sometimes altered, often 
missing, and always contained reference to the upload directories, 
which no longer existed. 



From Confusion and Chaos to Clarity and Hope 341 



Tom S. Bird song paper 
u0003_0000159 Manifest Page 

Administrative Information 

• u0003 0000159.vl.xml 

Collection Level Metadata 

• u0003 0000159.vl.txt 
Content 

• u0003 0000159 0000001 0001.tif 

• u0003 0000159 0000001 0002.tif 

• u0003 0000159 0000001 0003.tif 

• u0003 0000159 0000001 0004.tif 

• u0003 0000159 0000001 0005.tif 

• u0003 0000159 0000001 0006.tif 

• u0003 0000159 0000001 0007.tif 

• u0003 0000159 0000001 000S.tif 
■ U0003 0000159 0000001 0009.tif 

• u0003 0000159 0000001 0010.tif 

• u0003 0000159 0000001 OOll.tif 

• u0003 0000159 0000001 0012.tif 

• u0003 0000159 0000001 0013.tif 

• u0003 0000159 0000001 0014.tif 

LOCKSS system has permission to collect, preserve, and serve this 
Archival Unit 



Figure ALAB-4 University of Alabama Libraries LOCKSS Manifest example. (© 2009, 
University of Alabama Libraries. Used with permission.) 

In the metadata, if indeed there was an identifier, it was stored in 
multiple different fields in different forms. Sometimes there was 
nothing at all to indicate what the original file name had been. 



342 From Confusion and Chaos to Clarity and Hope 

We studied the file naming schemes used in all the 30-some 
existing collections, working out how to rename the files to retain the 
ordering of delivery, the intended organization of complex files, and 
the relationships between related items. Analyzing the metadata in the 
CONTENTdm (OCLC, 2009a) database, we matched up what 
information we could locate with what little had been retained with 
the archival files, and slowly filled in the gaps. Whatever sorting and 
matching could not be scripted was done by hand, often requiring 
research and investigation. 

During this time we explored the use of 7train (Fogel & Hetzner, 
n.d.) for transforming exported CONTENTdm Standard XML 
metadata into METS (Library of Congress, 2009b) files, which 
reordered and clarified the relationships between complex objects. 
Based on grain's method of using the first Dublin Core (DCMI, 2009) 
identifier field in the export as the file name for the resultant METS 
file, we selected this field for our file identifier, and tagged it the same 
in every collection. Our metadata librarian (Mary Alexander) has 
worked hard to remediate the metadata, entering the correct unique 
file identifier in the specified location in all records. Only by repairing 
the descriptive metadata to consistently reference the correct 
identifier, can we match our exported metadata to the appropriate 
archival files. 

The benefits of using this process are that the Dublin Core 
metadata assigned to the object at the top level is retained, the 
organization of the complex digital object is retained, and the 
transcripts or Optical Character Recognition (OCR) content are 
included in the resultant METS file. However, what is lost includes 
any metadata which does not map to Dublin Core, value-added labels 
which clarify the content in the fields, and page-level metadata beyond 
the title and transcript. If no archival file was used in upload, it is not 
referenced in the METS; those referenced may have their file name 
altered, and always refer to the upload directory. If the location of this 
directory is not corrected after upload to reference the actual location 
of the archival file, this is useless for reuniting metadata and archival 
content. 



From Confusion and Chaos to Clarity and Hope 343 

This version of METS was designed for web delivery, not for 
preservation. California Digital Library created 7train to enable 
repositories in the state that were using CONTENTdm to participate 
in their state-wide digital federated search service (Fogel, 2006), 
which requires METS (California Digital Library, 2009). The METS 
file contains links into the CONTENTdm software for access to the 
thumbnails and service derivatives. For preservation, these links will 
be useless, as delivery systems change rapidly, if indeed the content is 
still online. Additionally, the /train METS contains no technical or 
preservation metadata, as none was created by, or exported from, 
CONTENTdm. 

Seeking to make our METS files more useful, we determined how 
to add technical metadata, and analyzed the database structure and 
storage directories to identify the actual location and name of 
thumbnails and service images. Scripting to replace the derivative 
links and archival file reference with full path links to the actual files 
proved to be more trouble than it was worth. 

However, as our storage system began taking shape, another 
option emerged. Our storage structure reflects the compound file 
structure, creating an unambiguous arrangement which will survive 
any delivery software. A digital archivist of the future should have no 
trouble reconstructing our content. The METS file itself may be 
redundant. Rather than altering the 7Train METS file to meet our 
preservation needs, we decided instead to leverage our organizational 
scheme to meet the challenge. All we really need to do is to create the 
technical metadata, name it correctly, and drop it in the right 
directory (Figure ALAB-5). Then we will compile the metadata and 
content links for an item via script into a preservation-ready METS 
file for long-term storage. 



344 From Confusion and Chaos to Clarity and Hope 



T 13 U0003 






▼ 3 


0000001 




P 0000001 


T 


3 0000002 




T 


0 


0001 






T 


[3 Metadata 








1S] u0003_0000001_0000002_0001.mix.xml 








[Sj uOOO _ _O0O0OOl_O0OOOO2_O00 1 . a c r.txt 








S u0O0__O0O0OOl_O0OOOO2_O0Ol.tif 




T 




0002 






T 


p=_ _ 
pl Metadata 








H] u0003_0000001_0000002_0002.mix.xml 








@) uOOO 3_000000 1_0000002_0002 .oc r.txt 








ffi u0003_0000001_0000002_0002.tif 




T 


m 


0003 






T 


3 Metadata 








H] u0O03_O0O0OOl_O0OOOO2_O0O3.mix.xml 








13 u0003_0000001_0000002_000_.ocr.txt 








1 u0O03_O0O0OOl_O0OOOO2_O0O3.tif 




► 


3 


Documentation 




T 


3 


Metadata 








H u0003_0000001_0000002.mods.xml 


► 


0 0000003 


► 


0 0000004 


► 


0 Documentation 


► 


13 Metadata 



Figure ALAB-5. University of Alabama digital file organization for preservation. (© 
2009, University of Alabama Libraries. Used with permission.) 

Software Limitations on the Number of collections 

The expansion of Digital Services had a major impact on Hoole 
Special Collections. Archivists were scrambling to find sufficient 
content to digitize at first; small collections were easiest to pull 
together, so we suddenly were digitizing many tiny collections. What 
an archivist considers a collection is determined by provenance, not 
by quantity. 

However, CONTENTdm (version 4.3) was designed to support no 
more than 200 collections. It was clear that we could no longer define 
our collections the same way our delivery software did. Therefore we 
needed to reorganize our content in a way that provided the user 



From Confusion and Chaos to Clarity and Hope 345 



access desired by our archivists, while meeting the constraints of our 
current delivery system. 

We met with the archivists to ask what kind of grouping made 
sense to them. After much debate, they finally selected date ranges for 
most of our content, which corresponded to particular eventful 
periods in American History. Access to each of the digitized analog 
collections would be managed by canned links searching for the 
analog collection name in the Dublin Core relatiomisPartOf field. 
Archivists assigned each collection to a time period, and the metadata 
librarian and I began to sort out how to merge multiple collections 
and split combined ones. 

When merging collections in CONTENTdm, the two collections 
being merged must have identical metadata fields. It was during this 
process that we realized the depth and variety of our metadata across 
all those 30-odd collections. It is far more than can be captured in 
Dublin Core, and we realized that we wanted to retain the value-added 
labels. Knowing that a person is a photographer or a lyricist or a 
composer or a performer is far more valuable than can be conveyed by 
Dublin Core "creator." 

Our metadata librarian combined all our metadata fields into a 
single spreadsheet, so that all the CONTENTdm "containers" could 
handle any of our collections. Since each label requires a different 
spreadsheet column for ingest, we found we needed 87 columns. 
However, by hiding the columns that aren't needed, this becomes 
manageable. Different versions of this spreadsheet are used for each 
type of material, simply with different hidden columns. Since all the 
containers have the same metadata configuration, this streamlines 
uploads. The metadata librarian need only unhide the columns and 
export the spreadsheet for upload. Recognizing the value of the many 
tags in use, we decided to map all the fields to MODS (Library of 
Congress, 2009c) to discover if that metadata standard would be 
capable of retaining all our descriptive information; and indeed it 
does. 

My associate dean recognized the possibilities for further 
leveraging our current file organization for better user access and 



346 From Confusion and Chaos to Clarity and Hope 

delivery, sidestepping many of the restrictions placed on us by 
CONTENTdm. He procured the services of a talented programmer 
(Tonio Loewald), who proceeded to write a translator, which, given a 
template, can translate any Unicode tab-delimited spreadsheet into 
separate XML records of the desired schema. His Archivists Utility 
(Loewald, 2009b) reads in our 87-field spreadsheets and creates a 
MODS XML record from each line. This enables us to capture all of 
our metadata, not just the Dublin Core elements, outside of the 
delivery software for preservation storage. 

We added scripts using open-source software such as 
ImageMagick (ImageMagick Studio LLC, 2009), LAME (Cheng et al., 
2009), and Tesseract_OCR (Google, 2009) to generate derivatives 
from the archival files. These are placed into web-accessible 
directories that mirror our archive structure. By adding the MODS, 
and now our newly emerging EADs (Library of Congress. 2009a) we 
now have all the components for an open, modular delivery system. 
Under my associate dean's direction, the programmer has built an 
XML schema-agnostic delivery system, Acumen (Loewald, 2009b), 
that reads the XML where it lives in a live directory. Metadata and 
derivatives can be accessed easily and changed at any time without 
going through any software system. Web agents and web search 
engine crawlers can easily access our online content also, as it is not 
buried in a back-end database. Relationships between files are 
inferred by the file naming system, so that all components of an item 
are retrieved together, and all items in a collection can also be 
retrieved by using the collection identifier. In addition, since the file 
name communicates the hierarchy and relationships of files, digital 
content can be reorganized according to work flow or even moved 
between servers while keeping unchanging URLs. 

We're very excited about the possibilities this modular 
methodology offers. By bringing digital content up to the level of the 
web, we are setting the stage for semantic web applications and the 
development of user-friendly tools for access and reuse of our content. 



From Confusion and Chaos to Clarity and Hope 347 



Staffing Gap for Creation of Item-level Metadata 

Shortage in funding support for archivists is widespread. With 
the current attention to digitization of archival content, and the lack of 
funding for archivists, a critical gap is created. The archivists are well- 
versed in the knowledge of the time periods, issues, relationships, and 
people related to the archival content we seek to digitize; they are the 
personnel best suited to describe the material in a way which will 
enable users the best possible access. A knowledgeable archivist will 
be able to identify important personages of the time, particular 
buildings and locations, and can provide biographical, historical, and 
cultural context which gives meaning to the documents we seek to 
digitize. Since the metadata provides the necessary information for 
successful retrieval, correct descriptive information may be the most 
valuable knowledge capture we could provide. However, the archivists 
are hard pressed to provide even minimal processing to the 
burgeoning mass of incoming content. In order for us to move forward 
in digitization without more funding for archivists, we began to devise 
pilot projects to seek out alternative possibilities. 

Our first pilot project involved students creating item-level 
descriptive metadata as they digitized content. The collections chosen 
were small image collections containing a little over 200 photographs. 
To test for consistency, we assigned only one student to one of the 
collections, and four different students worked on the second 
collection. Within a few days it was clear that our careful instructions 
to the lone student were clearly insufficient. Her metadata 
spreadsheet was rife with errors; the primary focal person in the 
collection had his name misspelled seven different ways, and many 
words were abbreviated (and usually incorrectly). In 56 entries, we 
located 217 spelling and abbreviation errors. Grammar, punctuation, 
capitalization and spelling errors were abundant, and the captions and 
descriptions created were vague and unhelpful. 

In the jointly described collection, each student had his own 
methodology and focus, and the variations between the choices for 
descriptions and captions were sufficient to impair search and 
retrieval even within the collection itself. Where one student might 



348 From Confusion and Chaos to Clarity and Hope 

use the terms "infant" and "woman," a second student would use the 
terms "baby" and "lady." While the errors and misspellings were far 
less frequent, it was clear that for consistency, we needed either 
stricter guidelines and controlled vocabularies, or to limit the number 
of employees assigned to create metadata for any one collection. 
Additionally, since the students had no background information on 
the context of the collection, the time period, or the content, they were 
unable to identify well-known buildings, well-known personages, or 
other significant content of interest. This point became extremely 
clear when our students could not identify such historically important 
figures as Governor George Wallace or the locally beloved Paul "Bear" 
Bryant. 

A second pilot project involves the reuse of existing MARC 
metadata for sheet music which had been cataloged over a period of 
years by various catalogers. After extraction of the MARC records and 
analysis, we were dismayed at the variations in how the MARC fields 
were used. For example, we found the first line of text in fields 590, 
500, 740, and 246, along with other types of content. The arranger, 
lyricist, musician, and composer names were interspersed in field 
245c with various textual prefixes, preventing a systematic method of 
separation. Without remediation, transformation to another metadata 
scheme would not be recommended, as we cannot safely crosswalk 
these fields. Metadata librarians and catalogers are currently involved 
in repairs to the original records. 

A third pilot project involves having an archivist and metadata 
librarian each developing their own version of item-level metadata for 
the same photograph and correspondence collections. Each worker 
gathered time measurements, and neither party was allowed to view 
the others' descriptions until finalized. Both versions of each 
collection will be put online in the same web interface, and usability 
tests will determine to what extent the differences impact user 
experience. Questions for users will be derived from the known 
metadata in the finding aid, since the information in the series 
description would lead the user to expect success in searching for the 
item-level material that fit the description. If, for example, the series 
description mentions letters about boarding schools prior to 1900, a 



From Confusion and Chaos to Clarity and Hope 349 



query in the usability study would be to locate a letter about boarding 
schools written prior to 1900. This method of creating queries was not 
communicated to the metadata creators, so it could not skew the 
results. It is possible that neither metadata version will provide 
helpful results to the queries. In the final analysis, the level of usability 
for each version will be weighed against the time cost and availability 
of the personnel in the department that created it. We hope to be able 
to relieve the archivists of the burden of item-level metadata creation. 
While archivists know more about the content, and provide more 
detailed and informed metadata, their available time is very limited. 
Over the years, we have assumed that the expertise of the metadata 
creators is of primary importance, but new findings by Paul Conway 
(Conway, 2008) have thrown doubt on that belief. Since "the proof is 
in the pudding," we will determine if it is to our advantage to reduce 
the quality of metadata in order to get more of our content online. 

A fourth pilot project will depend solely upon EAD finding aid 
descriptions at the series and sub series level to provide findability 
and context for digitized items. Our Archival Access Coordinator 
(Donnelly Lancaster Walton) suggested that we seek to recreate online 
the experience of a patron exploring material in the reading room: a 
folder is opened, and the patron goes through the documents one at a 
time. No other information is available to describe the material. This 
methodology will enable archivists to focus on EAD finding aid 
descriptions, and our digitization team will take box after box and 
simply digitize content in order for web delivery. As content is 
digitized, links will be added to the online finding aids from the folder 
level, out to web directories which contain the digitized items, ordered 
as they were encountered in the folder. In this manner, we will be able 
to provide online access to huge collections for which we have 
insufficient resources to provide item-level descriptions. 

We have just been funded by NHPRC to demonstrate this low- 
cost, scalable model (University of Alabama Libraries, 2010a). Already 
we have developed scripts (University of Alabama Libraries, 2010b) to 
add links into the EADs and create minimal MODS records as 
quantities of scans become available. The software we develop to 
support this mass digitization method will be made available open 



350 From Confusion and Chaos to Clarity and Hope 

source. As soon as the content begins to appear online, we will 
conduct usability studies to compare the user experience between 
accessing content via the finding aid, versus item-level search and 
retrieval, and will publicize the results. 

The fifth pilot project we expect to undertake in the coming 
months will involve users in tagging our materials online. To the 
extent possible, we will repurpose available open source software, 
capturing the tags and free-text descriptions in a database for vetting 
by our metadata librarians. Tags may be made automatically visible in 
the web interface and included in indexing for search and retrieval. 
After review, acceptable entries will be added to the to the descriptive 
metadata record on file. Our first foray into this venue will be with 
photograph collections for which we have almost no descriptive 
metadata. Hence, any apparently valid contributions will be accepted. 
In this manner we also hope to build user interest and support as we 
build interaction into our interface. 

Summary 

In the real world, digitization support within an institution may be 
fragmented across several departments, only partially funded by 
donor gifts, and may suffer seemingly arbitrary restraints imposed by 
delivery software. Some of the departments upon which success 
depends may be understaffed and unable to meet the demands for 
digitization support. In addition, most digitization is undertaken with 
little thought for the future, either in terms of transitioning to 
alternative delivery software, or long-term access to digital content. 

Obstacles are opportunities for creative solutions. We leveraged 
expertise available across the library to build a successful cross- 
departmental collaboration. We are developing open source software 
support for an open, scalable, modular digital content delivery system 
with consideration for long-term preservation. Our file organization 
patterns alone may prove to be a life raft for digital content which 
funding cuts have left unsupported. This or a similar patterning will 
make reconstruction of digital content far easier. By possibly adding a 
Baglt manifest (Boyko, Kunze, Littman, Madden & Vargas, 2009), it 



From Confusion and Chaos to Clarity and Hope 351 



becomes potentially feasible to zip up entire digital archives for long 
term storage in a safe repository or LOCKSS until such time as 
funding support returns. 

The delivery system we are developing will be free and open 
source, requiring minimal technical expertise. An underfunded 
institution will be able to use it to raise their content to the level of the 
web where search engines can promote their materials and web agents 
can provide greater usability. Since the actual metadata and delivery 
content is not ingested, but left in the web directory, it can be changed 
as needed, without risk of harm to the online delivery. Metadata 
schemes change regularly, and this XML schema-agnostic, modular 
solution is a low-cost, scalable, simple approach to building online 
digital libraries with the future in mind. 

Thus far, we still use CONTENTdm for participation in multi-site 
search systems that depend upon this software, such as our state-wide 
Alabama Mosaic (Network of Alabama Academic Libraries, n.d.). 
However, the winds of change are blowing. During an open discussion 
at the 2009 AlabamaMosaic Annual Meeting in Montgomery, 
Alabama, the director reported a conversation with an OCLC 
technician who stated that FirstSearch would replace the 
CONTENTdm Multi-Site server. Thus, continued support of the 
CONTENTdm Multi-Site Server (OCLC, 2009c) is suspect, as OCLC 
moves toward FirstSearch (OCLC, 2009b) for cross-database search 
support. Many small collaborative digitization efforts depend upon 
CONTENTdm, but they need not be held hostage by a proprietary 
system. Acumen could potentially fill the gap, offering a low cost 
option providing much of the same functionality without the 
constraints. 

We continue to explore how best to fill or reduce the gap created 
by the apparent need for item-level metadata for access and retrieval, 
and hope that the solutions we adopt in that area will help others as 
well. We have certainly found that working together with mutual 
respect and consideration has brought both challenges and 
unexpected benefits. By bringing everyone to the table, struggling 
together with the chaos before us and facing the same goals, we have 



35 2 From Confusion and Chaos to Clarity and Hope 

unleashed passion and creativity, enabling us to make astounding 
strides into clarity, organization, and hope for the future. 

References 

Boyko, A., Kunze, J., Littman, J., Madden, L. & Vargas, B. (2009). The 
Baglt file packaging format (V0.96). Retrieved November 14, 
2009, from http://www.cdlib.org/inside/diglib/bagit/ 
bagitspec.html 

California Digital Library. (2009). Online Archive of California 
(OAC). Retrieved November 15, 2009, from http://www.cdlib.org 
/inside/projects/oac/ 

Cheng, M., Taylor, M., Hegemann, R., Leidinger, A., Tominaga, T, 
Shibata, N, et al. (2009). The LAME project. Retrieved November 

14, 2009, from http://lame.sourceforge.net/index.php 

Conway, P. (2009). The image and the expert user: a qualitative 
investigation of decision-making. Paper presented at Archiving 
2009, Arlington VA. In Archiving 2009, Vol. 6. (pp. 142-150). 
Society for Imaging Sciences and Technology. 

DCMI. (2009). Dublin Core metadata initiative. Retrieved November 

15, 2009, from http://dublincore.org/ 

Fogel, P. (2006). CDL 7train Profile - CONTENTdm simple and 
complex objects in METS Metadata encoding and transmission 
standard. Retrieved November 15, 2009, from 
http : / / www.loc.gov/ standards / mets / profiles /00000010 .html 

Fogel, P. & Hetzner, E. (n.d.). 7train METS Generation Tool. 
Copyright the University of California Regents. Retrieved 
November 15, 2009, from http://seventrain.sourceforge.net/ 

Google. (2009). Tesseract-ocr. Retrieved November 14, 2009, from 
http:/ / code.google.eom/p/tesseract-ocr/ 

ImageMagick Studio LLC. (2009). ImageMagick. Retrieved 
November 14, 2009 from http://www.imagemagick.org/ 
script/index.php 



From Confusion and Chaos to Clarity and Hope 353 



Library of Congress. (2009a). EAD Encoded archival description 
(Version 2002). Retrieved November 14, 2009, from 
http://www.loc.gov/ead 

Library of Congress. (2009b). METS Metadata encoding & 
transmission standard. Retrieved November 14, 2009, from 
http : / / www.loc.gov/ standards / mets / 

Library of Congress. (2009c). MODS metadata object description 
schema. Retrieved November 14, 2009, from http://www.loc.gov/ 
standards/mods/ 

LOCKSS. (2008). What is the LOCKSS program? Retrieved 

November 14, 2009, from http://www.lockss.org/lockss/Home 

Loewald, T. (2009a). Acumen. Retrieved November 15, 2009, from 
http://acumen.lib.ua.edu/ 

Loewald, T. (2009b). Archivists utility. Retrieved November 15, 
2009, from http://lb-4i6-oo3.1ib.ua-net.ua.edu/notes/?f= 
Archivist%2oUtility.txt 

Network of Alabama Academic Libraries, (n.d.). Alabama Mosaic. 
Retrieved November 14, 2009, from 
http://www.alabamamosaic.org/ 

Network of Alabama Academic Libraries. (2009). The Alabama 
Digital Preservation Network (ADPNet). Retrieved November 14, 
2009, from http://www.adpn.org/ 

OCLC. (2009a). CONTENTdm Digital collection management 
software. Retrieved November 15, 2009, from 
http://www.contentdm.com 

OCLC. (2009b). FirstSearch Online reference. Retrieved November 
15, 2009, from http://www.oclc.org/firstsearch/ 

OCLC. (2009c). Multi-site server. Retrieved November 15, 2009, from 
http://www.oclc.org/firstsearch/ 

University of Alabama Libraries. (2010a). Septimus D. Cabaniss 
Papers digitization project. Retrieved March 9, 2010 from 
http://www.lib.ua.edu/libraries/hoole/cabaniss 



354 From Confusion and Chaos to Clarity and Hope 

University of Alabama Libraries. (2010b). UA libraries digital 
services planning and documentation. Retrieved March 9, 2010 
from http://www.lib.ua.edu/wiki/digcoll 

W3C. (2005). XML:id Version 1.0. Retrieved November 15, 2009, 
from http://www.w3.0rg/TR/xml-id/ 



Digitizing Colorado State University's 
Historic Photograph Collection: A Case 
Study 

Beth Oehlerts (Colorado State University Libraries) 



Abstract 

This chapter discusses the experiences of creating a medium-sized 
digital collection of the earliest photographs from Colorado State 
University's Historic Photograph Collection. The entire collection of 
500,000 photographs chronicles the history of Colorado State 
University, the city of Fort Collins, and Rocky Mountain National 
Park. Digitizing the first phase of this vast collection included 
capturing 5500 images from glass plate negatives, 750 images from 
magic lantern slides, and 7500 images from gelatin nitrate prints. 
These formats were chosen because they are the oldest and most 
fragile. The project was not without challenges, including utilizing 
untrained staff and students, coping with equipment problems, and 
creating avenues of communication to more than 40 people involved 
in the creation of the collection. This chapter will discuss these 
challenges and how we worked to resolve them. 

Keywords: Archives, Digital collaboration, Digital collections, Digital 
projects, Digital scanners, Gelatin nitrate prints, Glass plate negatives, 
Magic lantern slides, Photograph collections. 



355 



356 Digitizing Colorado State University's Historic Photograph Collection 



Introduction 

Beginning in the late 1990s, Colorado State University Libraries 
(CSUL) started large-scale digitization activities in partnership with 
the Colorado Digitization Project (CDP). The first of these was the 
Warren and Genevieve Garst Photographic Collection of wild animal 
photographs donated to CSUL and digitized with CDP funding 
(http://lib.colostate.edu/wildlife/). The Sidney Heitman Germans from 
Russia Collection was another early digitization project 
(http://lib.colostate.edu/gfr/index.html) funded by the CDP. These 
projects, and a series of smaller ad hoc projects over the next three 
years, were created following the CDP's best practices guidelines and 
gave a few staff the opportunity to learn digitization. 

In 2006 CSUL received the University Historic Photograph 
Collection (UHPC) a collection of 500,000 photographs documenting 
the history of Colorado State University, Fort Collins, and Rocky 
Mountain National Park. The collection came from the University's 
Office of Instructional Services and was given to CSUL's Archives and 
Special Collections Department with the understanding that it would 
be preserved and access given to a wider audience through 
digitization. Planning for the digitization of the first phase began in 
early 2007, as did creation of a finding aid for the earliest images in 
the collection, which included glass plate negatives, magic lantern 
slides, and gelatin nitrate prints from cellulose nitrate negatives. 
These photographs are also the most fragile materials in the 
collection. The goal was to digitize approximately 5,500 glass plate 
negatives, 750 lantern slides, and 7500 gelatin prints. 

Although we had no budget line dedicated solely for digitization, 
our work did have the support of both the Colorado State University 
(CSU) and CSUL's administrative teams. Funds were provided by the 
Libraries' administrative team to purchase digital scanners, including 
an archival-quality overhead scanner used for later archival 
digitization projects. Administrative support provided extra funding 
in the project when we needed to hire students and acquire additional 
server space to house our digital master files. 



Digitizing Colorado State University's Historic Photograph Collection 357 



Literature Review 

Of the recent literature addressing digitization, the focus is often on 
the technical and preservation-related issues that arise following the 
creation of a digital collection; less of the recent literature focuses on 
the staffing issues encountered when creating digital collections. 
Boock and Vondracek (2006) found that when institutions begin 
digitization activities they often add these tasks to staff 
responsibilities rather than hire new staff dedicated to digitization. In 
a 2005 survey of ARL libraries, they found a majority of the 
responding institutions capitalized on the existing knowledge and 
skills of staff, encouraging current positions to evolve rather than 
hiring new digitization staff, as was the case at CSUL. Boock continues 
the discussion, focusing on how the Oregon State University Libraries 
(OSUL) reassigned several staff positions to a new digitization 
production unit, initially using staff for scanning, quality control 
reviews, and metadata creation using a metadata schema (Boock, 
2008). OSUL relies on student employees for almost all of its digital 
imaging and metadata assignment, performing no quality control 
reviews on their work. In examining the organizational implications of 
digitizing, Sutton (2004) recalls how early digitization efforts were 
seen as temporary endeavors, requiring the temporary re-assignment 
of staff. 

D'Andrea and Martin (2001), reporting on digitization workflows 
at Temple University, discuss utilizing part-time student staff to 
supplement the work of digitization staff. DAndrea and Martin note, 
and our experiences confirm, there are positives and negatives in 
hiring students to work on digitization projects. The authors 
recommend hiring students whose interests match the project, not 
just those who apply for the job, and fully informing students about 
the nature of the work. 

Establishing Best Practices 



In 2005, in response to the growing need for local standards and 
procedures to guide the creation of CSUL's digital collections, CSUL 
created the Digitization of Local Collections Task Force and charged it 



358 Digitizing Colorado State University's Historic Photograph Collection 

to review existing national and international standards and best 
practices, and to recommend the steps necessary in building high- 
quality local digital collections. Among its findings, as reported in the 
final report, Digitization of Local Collections (Digitization of Local 
Collections Task Force, 2005), the Task Force identified a lack of 
formal project management, a need to prioritize projects, and the 
need to determine the impact in-house digitization would have on 
staffing. Based on the Task Force's final report and the Western States 
Digital Imaging Best Practices, version 1.0 (Western States Digital 
Standards Group, 2003), CSUL then created a series of local best 
practices covering all aspects of the digitization process, starting with 
guidelines on how to handle and care for materials to be scanned. Our 
policy is to capture an image that is a true representation of the 
original. To do this all the objects in the UHPC digitization project 
were digitized as archival objects, with no alteration made to the 
master image. 

With some of the glass plate negatives, lantern slides, and gelatin 
nitrate prints more than 100 years old, including glass plates that are 
cracked or chipped, we scan an object once to reduce the chance of 
damaging fragile materials by repeated handling. Master images are 
created as TIFF (Tagged Image File Format) files. The glass plates 
were scanned at 1200 dpi, the magic lantern slides at 2400 dpi, and 
the gelatin nitrate prints at 1200 dpi. The master images were 
deposited in a project folder on a central server prior to the 
application of metadata and ingest into our digital repository. 

Our best practices documentation, including national and 
international digitization standards documentation resides on the 
CSUL staff wiki where it can be accessed by all employees. The wiki 
includes a wide range of information related to creating CSUL digital 
collections, including how to operate the digital scanners by format 
and instructions for creating project-tracking Excel spreadsheets that 
record the work of each staff member. The information gathered on 
these spreadsheets enables the supervisor to monitor and evaluate the 
work of the scanning staff, identify inefficiencies, and keep the project 
on schedule. Workflow analysis helped support our proposals to hire 
additional student scanning staff, purchase additional scanning 



Digitizing Colorado State University's Historic Photograph Collection 359 

equipment, and identify changes in workflow that will improve future 
digitization projects. 

Project Planning 

UHPC digitization project planning began in early 2007, with 
discussions and project meetings between the project archivist, the 
digitization librarian, and the metadata librarian. The project's 
participants knew what would be digitized, what metadata was 
needed, and where the collection would reside, but we had to develop 
a road map to help us achieve these goals. 

That road map began to take shape with the UHPC finding aid, 
created by the UHPC project archivist. The digital projects librarian 
created the UHPC digitization project-tracking documentation and 
identified staff for scanning, created staff scanning schedules, and 
arranged for staff access to the Archives Annex where the materials 
were housed. The metadata librarian, working with the project 
archivist, developed a plan to use the information in the finding aid 
for the metadata. Then the digital services librarian developed the 
procedures for loading them into our content management system. 

In September 2007, the UHPC digitization project scanning 
started with a small group of test scans to determine the best 
procedures for creating the digital images. Using CSUL's standards- 
based digitization guidelines, staff created approximately one dozen 
images. These scans were reviewed for accuracy and quality by both 
the project archivist and the scanning supervisor; when the pilot 
images met the requirements for the project, scanning began. 

Utilizing Untrained Staff 

Initially the number of trained digital scanning staff at CSUL was 
relatively small, with a Library Technician I and a Library Technician 
II, who worked on earlier digitization projects and had experience 
with scanning photographs and text documents, and a Library 
Technician II, with limited scanning experience. A Library Technician 
III supervised the work. These four staff also had other job 
assignments: the Library Technician I assisted in preparing materials 



360 Digitizing Colorado State University's Historic Photograph Collection 

for the bindery and assisted with a large inventory project; and the 
Library Technician lis performed database maintenance tasks and 
also assisted with the inventory project. The Library Technician III, 
who reports directly to the digital projects librarian, supervises 
CSUL's in-house digital scanning, the database maintenance unit and 
the inventory project. She is responsible for creating and adjusting 
scanning staff schedules, monitoring scanning progress, contacting 
Library Technology Services (LTS) staff to maintain and repair 
scanning equipment, and performing quality control reviews of the 
scanned images. 

The project began with 56 hours of scanning per week. Unable to 
hire additional staff, we examined staff duties to identify tasks that 
could be streamlined or eliminated altogether. A decrease in the 
number of print journal subscriptions and a reduction in the number 
of titles sent to a commercial bindery meant those saved hours could 
be applied to digitization. The same was true with tasks associated 
with database maintenance. For example, when approached by the 
CSUL Government Documents Unit to transfer the government 
documents database maintenance tasks to the Government 
Documents unit, we agreed and made the change. There were also 
some basic database maintenance tasks that could be performed by 
students, saving more time for digital activities. 

We identified one of our Library Technician lis whose knowledge 
of and experience with scanning, combined with her skills as an 
instructor, made her an excellent teacher. With these skills she trained 
our staff to create the high quality digital images we needed and now 
trains all our scanning staff and students. 

Experience has shown us that the level of training needed for 
creating digital images varies by individual and must be tailored to the 
individual. Each person brings various levels of experience and 
understanding to digital scanning. Aside from the level of experience, 
the unique nature of the materials, coupled with unfamiliar models of 
equipment, requires individualized training. Training begins with 
scanner instruction and viewing Cornell University Library's Moving 
Theory Into Practice Digital Imaging Tutorial (Cornell University 



Digitizing Colorado State University's Historic Photograph Collection 361 

Library, 2003.) Our digital imaging trainer spends 2-3 weeks teaching 
the trainee the local, national, and international standards and best 
practices, how to operate all of the digital scanning equipment, how to 
handle the materials being digitized, and any necessary project- 
specific scanning requirements. It usually takes staff, with multiple 
job responsibilities, one year to acquire the skills to create high quality 
archival digital images. 

In December 2007, a couple of months into the project, the 
scanning was not progressing as quickly as expected. We hesitated to 
consider using student employees due to the fragile condition of the 
glass plate negatives and magic lantern slides, but there was no other 
option; the project needed to keep on schedule. We trained our most 
experienced and mature student employee first and eventually hired 
more students to assist with scanning. The first three hourly student 
scanners spent a combined average of 36 hours per week scanning. As 
summer approached we asked and received permission to offer 
summer employment to two of the students. The ability to keep them 
working during the summer, at a total of approximately 60 hours per 
week, helped us keep the project on schedule. 

Midway through the project, in the spring of 2008, four members 
of the copy cataloging staff joined the UHPC digitization project. They 
were trained to assist the metadata librarian with metadata creation, a 
natural progression of their duties. The copy catalogers spent 
approximately 20 hours per week throughout the project assisting 
with metadata. 

The UHPC project archivist, our authoritative resource for the 
project, was also our daily contact for the project. Creating images at 
the Annex proved to be a benefit for the scanning team because the 
project archivist's office is located at the Annex. She was available to 
answer questions about the extremely fragile materials and instruct 
staff on how they should be handled. At the start of digitization, the 
project archivist was responsible for performing quality control 
reviews of the images, but we soon discovered this arrangement was 
not feasible. The project archivist was still in the midst of processing 
the collection's 500,000 images, and the logical solution was to have 



362 Digitizing Colorado State University's Historic Photograph Collection 

the scanning supervisor review the scans. To reduce the number of 
possible errors, scanning staff perform a review of their own scans 
prior to the end of their work day. Staff are expected to correct any 
errors found before a review by the supervisor. If the supervisor 
discovers images that need to be rescanned those are given back to the 
person who created the image. Correcting scanning errors became a 
learning opportunity and the person responsible for the scanning 
error was responsible for rescanning the image. If the problems were a 
result of malfunctioning scanning equipment, quality control reviews 
were performed on 100% of the images until the problems were 
corrected. 

Additional changes in job responsibilities occurred in the spring 
of 2008 when we replaced most staff working on the large inventory 
project with student employees. The change gave us approximately 32 
additional hours per week to devote to the UHPC digitization project. 
To accommodate the increase in hours we met with the archivists and 
the Archives and Special Collections Coordinator to request an 
extension of the hours at the Annex. The Archives staff was 
understandably reluctant to have others in the Annex when an 
archivist is not present, but they agreed to adjust the schedules of staff 
in the Annex and to extend the hours. The increase in the rate of 
production made meeting the project deadline a more realistic goal. 

By July 2008 all of the glass plate negatives and magic lantern 
slides had been scanned and once the quality control reviews and all 
rescans were complete we notified the metadata librarian. She began 
the process of preparing the metadata for the copy catalogers. Upon 
completion, the metadata librarian reviewed their work and requested 
corrections where needed. The metadata librarian then notified the 
digital services librarian, who began the process of ingesting the 
images and metadata into the digital repository. 

During the life of the UHPC digitization project we lost and 
gained staff and student employees. The impact on the project was 
challenging at times. In the late spring and again in the fall of 2008, 
two Library Technician I staff undertook several weeks of intensive 
training to acquire the skills to fill void of created by the loss of two 



Digitizing Colorado State University's Historic Photograph Collection 363 



employees. The new staff were unfamiliar with digital scanners, they 
had little or no experience handling fragile photographic materials. 
Fortunately they quickly became proficient and by the end of 2008 we 
added 24 hours a week back to our scanning schedule. 

The final format to be digitized was the gelatin nitrate prints. 
Work on digitizing this format began in August 2008 when we hired 
eight work study students to perform the gelatin nitrate print 
scanning. We began the fall semester with 132 hours per week: 80 
hours of student scanning time, 32 hours of staff scanning, training, 
and quality control review, and 20 hours of staff metadata creation. 
We hoped the students would be able to work largely on their own 
after training. The scanning supervisor worked closely with them for 
the first couple of weeks; she then let them work on their own. A short 
while later, while performing quality control reviews of their work she 
noticed the students were making a large number of errors, many of 
them basic scanning errors. Our experiment to reduce the close 
supervision of the students was not a success. Distractions, or perhaps 
the lure of the holidays, resulted in a lack of concentration. 

Scanning the UHPC collection took place at a site across campus. 
Saving the images to the Libraries' server from this remote site, and 
searching the web at the same time, caused the computers and 
scanners to slow and occasionally created serious technical problems. 
We now block internet access at all our scanning stations and more 
closely supervise the students' work. 

There are advantages and disadvantages with using students to 
work on digitization projects. The advantages include lower labor 
costs, a more flexible workforce, and the opportunity to give students 
valuable skills and good work habits. One disadvantage is the 
substantial investment in teaching students the skills and knowledge 
needed to create high-quality digital images only to have them 
graduate after a year or two of employment. Due to the repetitive 
nature of the work, it is difficult for some students to concentrate on 
scanning images. Many times students are not able or willing to work 
during school breaks. In spite of the disadvantages, we found that the 
positives of employing students outweigh the negatives. 



364 Digitizing Colorado State University's Historic Photograph Collection 

It is difficult to firmly estimate the number of hours of scanning 
time you will need when planning a digitization project. To 
compensate for staff leave and the uncertainty of student employment 
during breaks in the academic year or during final examinations, it is 
a good idea to budget extra time for digitization. Project downtime 
may occur because of equipment problems. In our case, while the 
scanners were being repaired we had to bring the scanning to a halt 
and reassign staff and students to other projects at CSU's Morgan 
Library. In addition to the scheduling problems on the UHPC 
digitization project, staff scanning schedules had to include 15 
minutes to walk across campus to the Archives Annex, where the 
scanning was done. By using our project-tracking documentation we 
were able to assess the progress of the project at various points and 
make any adjustments needed to keep us on schedule. The willingness 
of everyone involved to adapt and remain flexible was a major factor 
in the success of the project. 

Overcoming Equipment Problems 

The project began with just two Microtek Artixscan i8oof flatbed 
scanners. The Microtek scanners were purchased prior to the project 
startup and were chosen for their design, which includes a drawer 
where the glass plate rests emulsion-side up without touching glass or 
the light source. This separation is desirable when scanning fragile 
glass plate negatives. The scanners were installed and calibrated by 
the LTS staff, the equipment experts. During the first four months of 
scanning we encountered numerous equipment problems, including 
having to recalibrate the scanners every day, sometimes several times 
a day to eliminate colored lines on the digitized images. 

Equipment problems plagued us through much of this project and 
adding staff to scanning team did not help the situation. With more 
people scanning the scanners were in operation for 50 or more hours 
per week. This high rate of production took a toll on all the scanning 
equipment and the scanners were showing the stress of overuse. We 
needed more scanning equipment for both technical reasons and to 
increase production. The digital projects librarian recommended staff 



Digitizing Colorado State University's Historic Photograph Collection 365 

operate two scanners at a time, as one way to increase production. 
With 20 minutes to scan one glass plate, staff had time to prepare one 
image for scanning while the other was being scanned. We decided to 
follow this recommendation and requested that CSUL purchase two 
Epson Expression 10000 XL scanners, a model that best suits our 
needs. The scanners arrived in January 2008 but then we had to wait. 
It took LTS a couple of weeks to install the scanners. At the same time 
one of our Microtek scanners malfunctioned. Attempts to repair the 
scanner were unsuccessful and the only option was to return it to the 
manufacturer for refurbishment. The equipment installation delays 
and failures caused us to suspend the UHPC scanning for two weeks. 
We had to reassign staff and students to other projects in the interim. 

By February 2008 the four scanners were in place and working, 
but then we were experiencing other problems. Our new Epson 
scanners were producing Newton's Rings, a series of concentric, 
alternating light and dark rings centered at the point of contact 
between the glass surface of the scanner and the glass plate negative, 
on the scanned images (Illueca, Vazquez, Hernandez, & Viqueira, 
2002). The older scanners were beginning to produce banding 
artifacts (straight lines) on the images. As this only happened with the 
glass plate negatives, we switched from scanning glass plate negatives 
to scanning the magic lantern slides until we could determine a 
solution to the problems. A resolution to the Newton's rings problem 
came from the digital projects librarian, who thought an acid-free 
paperboard (barrier board) frame, similar to mat boards used in 
picture framing, might resolve the problem. The frame raises the glass 
plate from the scanner glass just enough to allow air to circulate 
between the plates of glass and eliminates the Newton's rings. During 
the digitization of the University of North Carolina at Chapel Hill's 
William Blake Archive, scanners encountered Newton's Rings when 
creating images from transparencies. Their solution was to scan the 
images directly and not through glass (Viscomi, 2002). This solution 
was not practical for us. Using a frame to separate the glass plate 
negatives, which are extremely fragile, from the scanner glass 
provided a safe method to protect the emulsion side of the plates. The 



366 Digitizing Colorado State University's Historic Photograph Collection 

banding artifacts problem was resolved when we began a daily 
cleaning and re-calibrating of the scanners. 

There was also the challenge of creating quality images from over- 
exposed and under-exposed glass plate negatives. We adjusted the 
scanner to capture either a darker or lighter image to correct the poor 
exposure. One point of pride for us was that we did not break or 
damage any of the fragile glass plate negatives in the collection. 

Improving Interdepartmental Communication 

Prior to the UHPC digitization project most digitization work was 
small ad hoc projects that did not require planning meetings, project- 
tracking documentation, detailed condition assessment guidelines, or 
extensive quality control reviews. As a result, not everyone 
participating in creating digital projects was aware of a project's 
status. Our digitization procedures changed with digitizing the UHPC 
collection. During the digitization project not all of us worked in the 
same location. The project archivist worked in one building, the 
metadata librarian, the digital projects librarian, and the digital 
services librarian worked in another. We needed to develop a project 
management structure that could foster communication and promote 
collaboration. Project communication, which was taking place via 
email, often resulted in further 'follow-up' emails. We realized that 
everyone involved in the project needed to know what was happening 
and why. 

In early 2008 the Repositories Matrix Team, CSUL's 
administrative and policy-making team for digital activities, created 
the Digital Projects Management Plan Working Group (Working 
Group). The Working Group's charge is to monitor the progress of all 
digitization projects. It includes representatives from Archives and 
Special Collections, Metadata and Preservation Services, and Digital 
Repositories Services; every department participating directly in 
digitization. The Working Group began meeting weekly in February 
2008. The initial focus was on the issues and problems surrounding 
the UHPC project. Now members report on digitization project news, 
digital equipment problems and equipment purchases, staffing issues, 



Digitizing Colorado State University's Historic Photograph Collection 367 

and conclude with a 'round robin' of project status reports. All policy 
questions are forwarded to the CSUL Repository Matrix Team. 
Building on the Working Group model, we now schedule pre-project 
planning meetings for all new digitization projects. These meetings 
focus on the scope of the project, staffing needs, equipment needs, and 
the creation of a digitization timeline. Attendees include faculty and 
staff who will be involved in the creation of the digital collection. 

The collaboration that formed during the UHPC project 
contributed much toward improving interdepartmental 
communication. The project required a close working relationship 
along with frequent meetings and emails between the project 
archivist, the digital projects librarian, the metadata librarian, and the 
digital imaging staff. Group members were willing to share their 
knowledge and skill with others and remain flexible while workflows 
shifted and changed course during the life of the project. Those 
involved with building digital collections at CSUL have learned that 
building quality digital collections requires planning, good 
communication, and a commitment to a collaborative endeavor. 

Summary/Lessons Learned 

Digitization of the UHPC collection provided several lessons. Within 
months of commencing the project, with enough data to create an 
accurate picture of our progress, we realized that we needed more 
staff, more equipment, and more time to complete the project. While 
hiring additional digitization staff was not possible, our option was to 
train staff in other units to assist with portions of the work. As 
mentioned above, the willingness of everyone involved to remain 
flexible was a major factor in the success of the project. 

When budgeting time for a digitization project, experience taught 
us that 20% more time should be added to the timeline, to allow for 
problems of all types to be resolved. Developing a realistic project 
budget and project timeframe will also help administrators and others 
understand the true costs of building a valuable digital collection. A 
short pilot project, conducted prior to the start of the project, can 
reveal quite a lot. It will show you how much time it really takes to 



368 Digitizing Colorado State University's Historic Photograph Collection 

scan the materials; which scanner settings should be used for a 
particular format; and whether it is possible to make textual 
documents searchable. Although we did conduct a short pilot, it was 
designed to learn how to operate the scanning equipment. We should 
have gathered other information, such as how long it takes to scan a 
glass plate negative, because soon after the scanning began we 
realized that we had wildly underestimated the time we thought was 
needed to scan these negatives. The 5 minutes scheduled per scan in 
fact turned into 20 minutes. A more comprehensive pilot would have 
revealed, among other things, what was technically feasible. 

The UHPC digitization project underscored for us the importance 
of regularly sharing information with the project's participants. Good 
communication, as mentioned above, was vital to understanding the 
UHPC project's goals and unique digitization requirements. With 
more than 40 people involved in its creation, sharing information 
helped us understand how to handle the fragile archival materials, 
avoid many image rescans, understand the reasons for many of our 
equipment problems, create solutions, and keep on track to meet the 
project's deadlines. 

Conclusion 

By June 2009 the photographs in the first phase of the project were 
digitized, and the images and metadata began to be loaded into the 
digital repository. Each digitization project is unique, and each has the 
potential to present new challenges. Though it might not be possible 
to anticipate every outcome, there are some things that can be done to 
prepare the project for success. Obtain the support of your 
administration or funding agency for the project; develop good project 
planning skills and implement them; create tools for communicating 
with all project participants and document all the processes and 
workflows of the project. Subsequent digitization projects from this 
vast collection will benefit from what we learned during the first phase 
of the UHPC digitization. 



Digitizing Colorado State University's Historic Photograph Collection 369 



References 

Boock, M. (2008). Organizing for digitization at Oregon State 
University: A case study and comparison with ARL libraries. 
Journal of Academic Librarianship, 34, 445-451. 

Boock, M., & Vondracek, R. (2006). Organizing for digitization: A 
survey. Libraries and the Academy, 6, 197-217. 

Cornell University Library. (2003) Moving theory into practice 
digital imaging tutorial. Retrieved on March 31, 2010 from 
http : / / www. library. Cornell . edu/ preservation/ tutorial 

DAndrea, P. & Martin, K. (2001). Careful considerations: Planning 
and managing digitization projects. Collection Management, 26(3), 
15-28. 

Digitization of Local Collections Task Force. (2005). Digitization of 
local collections. Retrieved on March 31, 2010 from 
http://digitool.library.colostate.edu/ 

Illueca, C, Vasquez, C, Hernandez, C, & Viqueira, V. (2002). The 
use of Newton's rings for characterizing ophthalmic lenses. 
Ophthalmic and Physiological Optics, 18, 361-362. 

Sutton, S. (2004). Navigating the point of no return: Organizational 
implications of digitization in special collections. Libraries and the 
Academy, 4, 233-243. 

Viscomi. J. (2002). Digital facsimile: Reading the William Blake 
Archive. Computers and the Humanities, 36, 27-48. 

Western States Digital Standards Group. (2003). Western States 
digital imaging best practices. Version 1.0. Retrieved on March 31, 
2010 from http://www.bcr.org/dps/cdp/best/wsdibp_vl.pdf 



Entering the Digitization Universe: One 
Catalog Librarian's Experience at an 
Academic Library 

Mary Rose (Southern Illinois University Edwardsville) 



Abstract 

This chapter describes a catalog librarian's experience with an 
academic library's digital collection initiative. The author discusses 
how the library handled technical challenges and established policies 
and procedures during the process of creating its first digital 
collection. The effects of external pressures from consortial 
requirements and organizational change are also discussed. The 
author describes technical decisions specific to the first project and 
more general technical issues like customization decisions and 
decisions about filenaming convention. The processes involved in 
establishing selection criteria and rights and permissions policies are 
described. The author also provides a brief overview of three 
subsequent digital projects. The author concludes by speculating on 
how the library's digital presence will grow in the future. 

Keywords: Academic libraries, Catalog librarians, CONTENTdm, 
Digitization, Digital collections. 



Lovejoy Library at Southern Illinois University Edwardsville (SIUE) 
entered the universe of searchable digitized collections in 2008. We 
encountered several issues along the way to completing our seminal 
project. There were technical challenges to be met, and we had to 



370 



Entering the Digitization Universe 371 



establish procedures and policies. We also encountered external 
pressures due to our reliance upon consortial services and as a result 
of organizational changes at the University. This chapter is a narrative 
of this experience and a speculation about the future. 

Background: the preliminary steps toward establishing a 
digital initiative and vision 

In 2006-2007, Lovejoy Library administration took the first steps 
toward establishing a digital projects initiative by forming a 
CONTENTdm committee and acquiring access to CONTENTdm 
software as a member of the Consortium of Academic and Research 
Libraries in Illinois (CARLI). The software is installed and maintained 
on CARLI's server. Two Lovejoy staff members received training in 
the use of CONTENTdm; however, neither staff member was 
empowered with a mandate to create a digital collection. The initiative 
essentially stalled. When I joined the University as the Library's first 
catalog and metadata librarian in May of 2007, I recognized that 
getting Lovejoy fully engaged in the creation of digital collections was 
a main priority of the position. The aforementioned staff members 
immediately and gratefully handed their CONTENTdm workbooks 
over to me and notified the consortium that I was now the primary 
contact for coordinating the Library's use of this software. I had never 
previously used CONTENTdm but became intimately familiar with it 
over the course of the next several months. Lacking training or 
experience, I relied heavily on support services at CARLI to effectively 
leverage the software. I also took a generic metadata creation 
workshop and studied Dublin Core. 

I quickly became aware that two digitization projects were being 
spearheaded by two tenured faculty librarians as candidates for our 
initial digital collection: one somewhat aggressively as a grant project 
and the other more casually without the impetus of a grant. Being 
naive with regard to the politics of the organization, I deferred to 
others who decided to give precedence to the grant-funded project. 
The CONTENTdm committee subsequently decided that the Library 
needed a process for evaluating and prioritizing potential digital 
collections. Perhaps this was a response to the way in which resources 



372 Entering the Digitization Universe 

had been committed to the first project because of a schedule driven 
by external funding. Or perhaps it was the usual librarian caution that 
any new undertaking will grow to unmanageable proportions if fed too 
liberally. Perhaps the desire for oversight was motivated by 
recognition that the shape of our accumulated digital collections over 
time would define the character of the Library to a significant degree, 
and whether this was ad hoc or directed was not a matter of chance 
but of choice. Whatever the reason, a digitization selection 
subcommittee to the collection management committee was proposed 
by a tenured library faculty member at the first CONTENTdm 
committee meeting I attended. 

The digitization selection subcommittee became entwined with 
the Library's vision regarding digital initiatives. The subcommittee's 
charge was officially established as being the body responsible for 
receiving and evaluating digitization project proposals and making 
recommendations to the parent collection management committee 
regarding acceptance and prioritization of said proposals. The 
advisory group comprising the subcommittee included all of the 
library faculty administrators plus the Director of Development 
(essentially the marketing administrator) and the Director of 
Academic Computing. The subcommittee was rounded out by the 
Catalog and Metadata Librarian (me), the Electronic Resources 
Librarian, the Archivist and Special Collections Librarian (serving as 
chair), and whichever subject librarian was participating in a specific 
digitization proposal. The group resolved to create a proposal form to 
guide proponents in describing the subject, extent, rationale, funding, 
etc. of their project ideas. Selection would be accomplished by 
carefully evaluating the relevance of a project to the Library's mission 
and the advantages a digital platform was expected to provide for the 
particular included items, such as wider accessibility for heavily used 
resources, easier use of delicate or cumbersome materials, and 
improved access to text-rich content through electronic searchability. 
Selection criteria suggested by the Northeast Document Conservation 
Center were incorporated into the subcommittee's official position. 
The Center frames selection around three basic questions (Gertz, 
2007): 



Entering the Digitization Universe 373 



• Should [the materials] be digitized? Is the collection important 
enough, is there enough audience demand, and can sufficient 
value be added through digitization to make it worth the cost 
and effort? 

• May they be digitized? Does the institution have the 
intellectual property rights to permit legal creation and 
dissemination of a digital version? 

• Can they be digitized? Will digitization achieve the goals of the 
project, given the physical nature of the materials and their 
organization, arrangement, and description? Does the 
institution have the technical infrastructure and expertise to 
create digital files and make them available to users now and 
in the future? 

Challenges encountered during the first digital project 

Our pilot digital collection was the KMOX sheet music 
digitization project. Lovejoy Library's Music Special Collections 
includes a gift from KMOX of over 48,000 music titles compiled by 
the St. Louis-area radio station: the live studio orchestra's complete 
performing music library. The titles date from the early 1900s. A 
subset of this collection, identified as being published prior to 1923 
and hence in the public domain, became the target digital collection. 
Academic Computing, an entity under the administration of the 
Library's dean, scanned the sheet music in color at 600dpi, enlarged 
400% during scanning and saved as uncompressed tif files. I began 
working on the project in earnest in January of 2008, with my first 
real technical task being to understand CONTENTdm enough to 
design a structure to showcase the collection effectively. Eventually I 
settled on a strategy: Each piece of sheet music would be what is 
known in CONTENTdm terminology as a compound object. Metadata 
would be supplied at the object level, meaning each piece of sheet 
music would have its own metadata but the individual pages 
comprising a given title would not be described separately. 

The Fine Arts Librarian had obtained grant monies to hire 
graduate student assistants to help with the project. I trained the 
graduate assistants how to provide what catalogers consider 



374 Entering the Digitization Universe 

descriptive metadata. This is the metadata that is transcribed from the 
piece being described. In this case, descriptive metadata included the 
song title, first line of the refrain, and publication information. I also 
showed the students how to search the Library of Congress's free 
online authority file (Library of Congress, 2009) for authorized forms 
of names for the lyricists, composers, arrangers, performers, and/or 
illustrators credited on the pieces. Finally, I created standard notes for 
the student assistants to apply, such as "piano, vocal" for an 
instrumentation note, "One color (purple)" describing the cover art, 
and "Includes advertisements" as a miscellaneous note. I reviewed 
their work and completed the metadata with subject analysis, detailed 
cover art description, and additional notes. 

Learning how to use CONTENTdm required a tremendous 
amount of time and energy during this first project. The effort was 
amply rewarded, however, since the functionality provided by the 
software suited our application perfectly. The software supports batch 
population of a collection via tab-delimited files. This facilitated 
collaborative metadata creation, since the graduate student workers 
could create Excel spreadsheets with preliminary metadata for groups 
of titles and then pass them on to me to complete. I subsequently 
converted the spreadsheets into tab-delimited files and uploaded the 
metadata into CONTENTdm along with the corresponding images. 
The compound object structure, in which several images comprise one 
digital entity, elegantly matches the character of multi-page sheet 
music. The software also provides the means for creating index boxes, 
which enhance access to the content beyond full text searches of the 
metadata. We decided to use this functionality to create index boxes 
for composers, lyricists, and subjects for this project. 

As stated previously, Lovejoy Library's digital collections are 
created under the consortial umbrella, using CARLI's CONTENTdm 
server. CARLI's collection of member libraries' digital collections is 
OAI-harvestable, and CARLI provides a means for member libraries 
to obtain usage statistics. But with these advantages come some 
constraints. CARLI requires all of their hosted collections to contain 
certain metadata fields, including (among others) Rights and 
Language fields mapped to the corresponding Dublin Core elements 



Entering the Digitization Universe 375 



and a Collection field mapped to Relation. The Rights field 
requirement motivated our CONTENTdm committee to address the 
thorny issues of intellectual property more promptly than we might 
perhaps have otherwise; as it was we needed to formulate a policy 
before publishing the KMOX collection. This proved to be the 
committee's most important task. The consortium specified that the 
Rights field should identify the intellectual property rights status of 
the digital resources in the collection and provide direction for users 
to contact the owner. This field could also be used to inform users of 
fair use laws. The committee consulted with the University's legal 
counsel to develop a rights and permissions policy in conformance 
with these guidelines. The digital rights and permissions statement 
that eventually evolved through the committee's deliberations 
authorizes "fair use" of the digital resources, provides references 
describing the legal limits of fair use, specifies the form of attribution, 
and provides the means for applying for additional permissions 
(Lovejoy Library, 2009). 

SIUE is responsible for the resources comprising our digital 
collections. Before we begin a digital collection project, we need to 
establish our right to create these component digital resources. This 
can be accomplished by using source materials in the public domain, 
securing permission for digitization and publication from the owner of 
the source materials, or actually purchasing the right to digitize and 
publish source materials. However, the rights status of the source 
materials is not always unambiguous. For the KMOX sheet music 
project, items within the public domain were identified as such by 
having a copyright or publication date prior to 1923 printed on the 
item. However, as I completed the metadata I noticed that some of the 
covers exhibited images clearly indicating that they were created after 
that date. For instance, the cover of "Come to the Fair" featured a 
photograph of the Trylon, Perisphere, and Helicline at the 1939 New 
York World's Fair, despite the fact that the music bore a copyright 
date of 1917. The cover for "The World Is Waiting for the Sunrise" 
depicted singer Mary Ford, who was born in 1924. Other pieces of 
music included advertisements for songs displaying later copyright 
dates. The Fine Arts Librarian consulted with legal counsel about the 



376 Entering the Digitization Universe 



status of these items. It was decided that we could include these pieces 
of sheet music in the collection if we didn't provide access to the 
individual pages that were not in the public domain, an approach that 
wouldn't affect the usability of the music itself. 

Another consequence of using CARLI's server is that our digital 
collections are subject to CARLI's "look and feel" requirements for 
uniformity. The consortium allows very little flexibility, as it wants to 
maintain a consistent look between the collections of member 
organizations. For our first project, this was actually a blessing. 
Designing a branded image is a lengthy process requiring resources 
(graphic design talent and technological tools and adroitness) and 
research (complying with the look and feel requirements of the SIUE 
website as a whole). As it was, the decisions I presented before the 
CONTENTdm committee were straightforward and simple. I made 
some mockups featuring the school colors in various combinations in 
the permitted areas; the voting process was fairly painless. I worked 
with Academic Computing personnel to get an official logo that 
conformed to CARLI's size constraints with the exact color specified 
by SIUE marketing guidelines. 

Challenges encountered during subsequent digital projects 

My second digital collection experience, the digitized presentation 
of a Civil War diary, was achieved in collaboration with the Social 
Sciences Librarian and a temporary staff worker under her 
supervision who had transcribed the entire diary. I learned how to use 
the transcription function in CONTENTdm and worked with the staff 
worker to render the transcription she had created into files that 
CONTENTdm could manipulate, i.e. individual text files with file 
names matching the corresponding image files. 

In the spring 2009 semester, I was the instructor of record for a 
student's Senior Project course. The student, who had worked in a 
library for several years and was considering going to library school 
after completing his bachelor's degree, wanted to learn about Dublin 
Core metadata and digital collections. Together we designed a project 
for him to create a Civil War collection under my supervision using 
digitized letters and ephemera loaned to the Library by an emeritus 



Entering the Digitization Universe 377 



professor of history. After completing background readings and 
papers, the student spent about five hours a week at the Library. He 
collaborated with me to make metadata decisions and learned how to 
use the CONTENTdm software, successfully completing the project in 
a semester's time. Working on this collection revealed a shortcoming 
of CONTENTdm in the way it supports managing metadata for 
disparate types of materials within a single collection. This project, 
which my student ultimately named the American Civil War 
Collection, is comprised of three different types of digital entities: 
letters, military orders, and songsheets. Adequately describing all 
three required a total of 27 different metadata fields. CONTENTdm 
does not have the functionality to organize metadata separately into 
subsets determined by their relevancy to particular included objects. 
Metadata manipulation (mapping, editing, etc.) after uploading is 
performed in a single interface in which all the metadata fields are 
displayed together. Fortunately in our case the small overall size of the 
collection meant coping with this limitation wasn't prohibitively 
awkward. 

The same spring the library administration hired a second catalog 
and metadata librarian, and together we began work on a fourth 
digital collection. This project featured digital photographs of 
architectural artifacts designed by architect Louis H. Sullivan and 
owned by SIUE, accompanied by digitized historic photos of the 
buildings on which the ornaments originally appeared. We worked 
with the Fine Arts Librarian and her graduate assistant to plan the 
organization and presentation of the images and identify the metadata 
we wanted to include. The graduate assistant gathered the raw 
metadata which my colleague and I translated into controlled 
vocabularies. We used the Getty Art and Architectural Thesaurus 
(AAT) (Getty, n.d.) for terminology for the ornaments themselves, 
materials of construction, and types of buildings of origin, 
supplementing the latter with Library of Congress Subject Headings 
when we felt it would be helpful. We began populating the digital 
collection in April, a process that took four months due primarily to 
delays in obtaining some of the images and associated descriptions. 
Leveraging CONTENTdm to create a meaningful structure for objects 



378 Entering the Digitization Universe 



in this collection proved challenging. We ultimately decided upon 
what CONTENTdm calls a monograph structure. A CONTENTdm 
monograph is a compound object with hierarchical levels, analogous 
to chapters in a book. We organized each of our digital entities to have 
two subsets (chapters) of images: artifact images and building images. 
Users click on one of these headings to reveal the images in the next 
hierarchical level. Although this structure isn't inherently intuitive, we 
felt it was the best fit from among the options available in the 
CONTENTdm software. The structure works well when browsing the 
collection as a whole or via the index boxes we supplied for artifact 
type and building of origin, but we are less enthusiastic with how it 
translates into retrieval from keyword searches. CONTENTdm has 
options for customizing the retrieval display that address some of our 
concerns, but the fact that the software isolates document- and page- 
level metadata in the search and display customization functionalities 
prohibits us from achieving our ideal result. 

In summary, the four digital collections that Lovejoy has created 
to date using CONTENTdm are: 

1. KM OX Popular Sheet Music , comprised of 118 objects and 558 
jpg files. 

2. William R. Townsend Civil War Diary, comprised of 14 objects 
and 356 jpg files. 

3. American Civil War Collection, comprised of 9 objects and 40 
jpg files. 

4. Louis H. Sullivan Ornaments, comprised of 64 objects and 191 
jpg files. 

The completion of each project was marked by announcing its 
availability to the University community and adding a link to the 
library website. I also created a catalog record in OCLC for each 
collection, and added all four collections to the CONTENTdm 
Collection of Collections database (CONTENTdm, n.d.). 

Issues to be addressed in future projects 

Organizational change has provided a source of external pressure 
concurrent with and affecting the progress of our digital initiatives 
and priorities. Lovejoy Library's dean left near the end of 2007 after a 



Entering the Digitization Universe 379 



long tenure as both Dean of Library and Information Services and 
Associate Vice Chancellor for Information Technology. Academic 
Computing had reported to the Dean in his latter capacity. Upon the 
Dean's retirement, the Provost decided to change the organizational 
structure so that the new library Dean would not have this dual 
responsibility. Academic Computing merged with the Office of 
Information Technology Services and now shares with it a new 
reporting structure separate from library administration. The new 
system began in July 2008. The interim Director of Technical Services 
began exploring a team approach to digitization. His plan centered 
around two major initiatives: purchasing a large format scanner for 
the library and hiring a digital imaging specialist, which were 
accomplished in 2008-2009. Library digitization projects could 
consequently be created without relying on Academic Computing 
personnel to scan materials. However, all of the aforementioned 
projects were digitized by various people before the purchase of the 
library's large format scanner and subsequent hiring of the Digital 
Imaging Specialist. The team approach has not yet been developed for 
producing CONTENTdm digital collections. 

In fact, creation of the image files began prior to my involvement 
with each of the projects except the first. As a result, a filenaming 
convention was never established. We discussed filenaming for the 
first project, the KMOX sheet music collection. We decided to use a 
transparent method: The images were named using a combination of 
the song title, composer name, date, and page number. An example is 
ByTheLight_Edwards, Gus_1909_001.jpg. This approach doesn't 
support generalization to future projects. As I researched the issue 
further, I grew to prefer a more systematic approach to filenaming. 
This idea inspired me to create an Image ID field in the metadata for 
each sheet music title in the collection, which I populated with an 
alphanumeric collection-specific accession number. But the actual file 
names corresponding to the jpg files weren't included in the final 
metadata: an inadvertent oversight resulting from my inexperience 
with how CONTENTdm handles tab-delimited files. The problems 
with file names persisted for all four of the projects described 
previously in this chapter. The Digital Imaging Specialist is working 



380 Entering the Digitization Universe 



with me and the other catalog and metadata librarian to establish a 
convention that works with our scanning equipment defaults. We have 
decided to adopt a cross-collection systemized convention similar to 
that described in the Wisconsin Heritage Online Digital Imaging 
Guidelines: 

File names for digital masters and derivatives need to be 
established before the scanning process. Systematic file naming 
helps not only to manage the project, but also ensures system 
compatibility and interoperability. It is generally recommended to 
assign an eight-character file name and a three-character 
extension, e.g. aaoooooi.xxx. This is sometimes called 8.3. File 
names should adhere to some general requirements. They should 
be: 

• Unique and consistent 

• Alphanumeric (consist of only letters and numbers) 

• Lowercase 

• Free of spaces and tabs 

• Numbered sequentially using leading zeroes (i.e. 001, 002, 
003, not 1, 2, 3) 

The files can be named after an original source collection or per 
project, depending on the needs of the local institution. Up to four 
letters can represent the project abbreviation or original collection 
name, e.g. hf for Harrison Forman Collection or sccl for Shawano 
City-County Library. The remaining digits indicate a unique file 
number. This is often simply sequential numbers prefaced with 
leading zeros. For example, digital images from the Harrison 
Forman collection project are named hfoooooi, hfooooo2, etc. 
(p. 5-6) 

I reached the end of my digital collection backlog with the 
completion of Lovejoy's fourth CONTENTdm-based collection in 
September 2009. I subsequently met with some of my colleagues to 
brainstorm ideas for additional digital collections. The result was a 
fantastic array of proposals employing audio and video files, featuring 
collaboration with other local institutions, and creating scholarly 
research products on a digital platform. The proposals were presented 



Entering the Digitization Universe 381 



to the selection subcommittee. The constituency of the subcommittee 
had been modified to reflect the organizational changes described 
previously in this chapter. The Director of Academic Computing was 
no longer a part of the subcommittee and the new Digital Imaging 
Specialist had been added. Although the majority of the subcommittee 
greeted the new project proposals with enthusiasm, the role of the 
selection subcommittee is currently being reconsidered and thus the 
project approval process is on hold. 

We are planning to purchase a server in cooperation with our IT 
department. Not only will this relieve severe storage problems during 
digitization workflow, it will also give us the option to explore creating 
portal pages to our CARLI digital collections or to host some 
collections locally. The Digital Imaging Specialist has a graphic design 
background and is highly interested in exploring creative ways to 
showcase our collections. 

Sorting out the process of green-lighting digital projects and the 
graphic and technological design of locally-hosted portals and 
collections will doubtless incur long and passionate discussion. The 
committee-driven process that is the default for all decisions at 
Lovejoy is not a painless one. Consensus-seeking, while attractive in 
theory, is impractical in many ways. But it is the culture of this 
institution and I suspect the culture of many similar institutions as 
well. Some issues along the way to realizing our digital initiatives thus 
far have been thoroughly discussed and resolved with thoughtful 
regard for the future, and some were hastily addressed with the main 
goal of overcoming a stalling impediment. Some of the best ideas 
proved insufficiently nimble to adjust to unforeseen developments. 
Some of the bad seed sown in the interests of forward motion has yet 
to bear the anticipated troublesome crop. Regardless, we are moving 
forward into new kinds of projects with a sharper focus on who we are 
and how we want to present ourselves. 

References 

CONTENTdm Collection of Collections, (n.d.). Retrieved December 1, 
2009, from http://collections.contentdmdemo.com/ 



382 Entering the Digitization Universe 

Gertz, Janet. (2007). Preservation and selection for digitization. 
Retrieved December 10, 2009, from 
http : / / www.nedcc.org/ resources /leaflets / 6Reformatting/ 
o6PreservationAndSelection.php 

The J. Paul Getty Trust, (n.d.). Art & Architectural Thesaurus Online. 
Retrieved December 8, 2009, from http: //www. getty. 
edu/ research/ conducting_research/vocabularies/aat/ 

Library of Congress Authorities. (2009). Retrieved December 1, 
2009, from http://authorities.loc.gov/ 

Lovejoy Library. (2009). Digital rights and permissions. Retrieved 
December 1, 2009, from http://www.siue.edu/lovejoylibrary/ 
about/digital_rights_and_permission.shtml 

Wisconsin Heritage Online Digital Imaging Guidelines (Version 2) 
(2009, September). Received December 2, 2009, from Wisconsin 
Heritage Online Wisconsin Library Services. 



From Argentina to Zambia: Capturing the 
Digital A to Z's of a Child Art Collection 



Kathleen C. Lonbom, Milner Library (Illinois State University) 



Abstract 

The International Collection of Child Art, residing at Illinois State 
University's Milner Library, is a collection of artworks produced by 
children and adolescents across a range of cultures and time periods, 
primarily mid to late 20 th century. This chapter discusses the 
collection's background and its role as a culturally expansive primary 
source. Information is provided about the Library Services and 
Technology Act grant funding awarded through the Illinois State 
Library to support the digitization project, Imagine Illinois and 
Beyond: Celebrating Creativity Through the Eyes of Our Children! 
The benefits and challenges of the digital conversion are discussed. 
Finally the chapter will look at alternate methods of image access, 
specifically audio description, to facilitate information discovery for 
viewers with a print disability such as vision impairment. 

Keywords: Academic libraries, Art, Audio description, Children's 
art, Cultural heritage materials-digitization, Descriptive metadata, 
Grant funded projects, Illinois State Library, Institute of Museum and 
Library Services, Library materials-digitization, Metadata, Special 
collections, Visual resources. 



383 



384 From Argentina to Zambia 



Collection Background 

The International Collection of Child Art (ICCA) is a resource 
comprised of artworks created by children and adolescents from 
around the world. This cultural heritage collection reflects the visual 
expressions of young artists who capture themes from the fantastical 
to the familial and a myriad of themes that fall between. The collection 
was initiated more than forty years ago at Illinois State University 
(ISU), Normal, Illinois, and is now curated and administered by the 
University's library. This primary source includes over 8,600 
accessioned children's artworks, from approximately 58 countries and 
cultures. The collection celebrates the creativity and innovative work 
produced by children with a multicultural perspective. The resource 
serves multiple audiences including students, academic scholars from 
across disciplines, and a wider public interested in viewing, learning 
about, and appreciating the imaginative vision that shapes a child's 
creative pursuits. 

ISU's Milner Library acquired the ICCA in 2000 from the ISU 
School of Art, which inherited the resource following the closing of the 
University Museum in 1991. The collection had been stored in a 
classroom until the college could no longer accommodate it due to 
space constraints, lack of support staff, and less than optimal 
conditions for storing and maintaining this resource. 

Virtual Shift: The Digital Project's Genesis 

University libraries holding collections with parallel cultural and 
historical value have grappled with similar challenges and 
opportunities presented by the digital conversion of a unique 
resource. Questions, both philosophical and practical, invite those 
embarking on a digital project to entertain a range of considerations 
perhaps not previously attached to the physical collection. Digitization 
of the Eastern North Carolina Postcard Collection project members 
recognized the ambiguous nature of assigning subject headings to a 
stand-alone image, relatively free of contextual information. The 
postcards in this collection usually were accompanied by text caption, 
but did not always have context provided by a monograph (Dragon, 



From Argentina to Zambia 385 



2009) . Colorado State University's project to digitize the University 
Historic Photographic Collection emphasized the collaborative 
process of converting a historic resource under a controlled timeline. 
The conversion involved a variety of project partners learning to speak 
each other's professional language including archivists, metadata 
librarians, and digital project managers (Hunter, Legg & Oehlerts, 

2010) .). A Latin American political poster collection, part of the 
University of New Mexico Libraries Center for Southwest Research 
and Special Collections, speaks to digitally documenting the transient 
nature of a resource by preserving ephemera such as posters. Similar 
to children's art, posters are not typically created to last through time, 
but document a specific and often meaningful moment in time 
situated at the edge of societal mainstream perspective (Stephenson, 
2006). Clifford Lynch's discussion of digitizing cultural heritage 
materials comments broadly on the discovery aspect of placing 
resources in an open electronic environment where unexpected and at 
times serendipitous communities form around such a collection. 
Lynch posits it is the objective of a digital library, expansive in its 
capacity to "enable and facilitate implicit communication" to provide a 
construct for the community building that develops around a 
collection (Lynch, 2007). 

The move to consider digitization of the ICCA was rooted in 
curricular needs to accommodate a generation of users already vested 
in electronic access. By virtue of the resource's sheer size, aging 
condition, and location, physical access is limited. The collection of 
two dimensional artworks is stored in a multi-use university 
warehouse that also serves as the library's storage site for a collection 
of less frequently used volumes and is the home of University 
Archives. The warehouse is off campus and largely off limits to the 
public. Although slides of the artworks were created in the 1970s to 
support teaching and study of the collection, the analog format was no 
longer a viable option to support use of the collection in the 21 st 
century academic environment. Included in the collection are sixteen 
thematic traveling exhibits that have been displayed widely and have 
garnered attention for the resource. Several hundred of the slides, 
selected from the collection's traveling exhibits, were converted to 



386 From Argentina to Zambia 

digital format in 2003, but overall the condition and quality of the 
slides was questionable and each image required color correction. The 
conversion of this small sample to a digital format brought attention 
to the collection when the images were mounted on the ICCA website 
and provided the spark that ignited the pursuit of funding to digitize 
the collection. Art, art education, English, and children's literature 
faculty actively using the collection in classes were supportive 
advocates for moving the resource into a fully accessible digital 
format. 

Setting the stage for digital conversion in the fiscally constrained 
environment of a public university requires strategic use of limited 
funding for library resources and creative thinking to move forward 
with a digitization project. Milner Library began researching grant 
options to move forward with the digital conversion of the ICCA in 
response to faculty feedback. 

The value of a grant award is not limited to the funding awarded 
to support a project. Significant gains are gleaned from a successful 
external grant award including the opportunity to work outside your 
own institution with the awarding agency, related publicity, 
promotion, and recognition of a collection. The external recognition 
has the potential to bring additional funding as a project's status is 
heightened, providing a scaffolding effect on which to build and 
enhance the project and its outcomes. Additionally, faculty librarians 
working towards tenure and promotion are keenly aware of the 
professional distinction associated with the potential of a funded 
competitive external grant and the implications for building a record 
of scholarly and creative activity. 

Grant awards also provide resources to fund hiring student 
assistants who collaborate and contribute to a project's success. In an 
academic environment a grant funded library digitization project can 
advance a student's academic career by affording opportunities to 
work on the multiple facets afforded by such a project. Grant funded 
projects, such as the proposed digitization of the ICCA, offer a wide 
range of opportunities for students from a variety of academic 
disciplines, testing and developing students' strength as a researcher 



From Argentina to Zambia 387 



in art, art history, history, foreign language and culture. Students who 
have worked with the ICCA come from a variety of majors: art, art 
history, arts technology, English, and language. A student can also 
learn about time and project management skills and supporting the 
goals and objectives of a project through their own contributions. 

Writing a successful grant proposal reflects a commitment of 
time, thought, and energy while providing an opportunity to compete 
and be recognized through the external agency awarding the grants. 
The Illinois State Library's (ISL) Library Services and Technology Act 
(LSTA) grant program, using funding made possible by a grant from 
the Institute of Museum and Library Services (IMLS), appeared to be 
a fitting choice for a proposal submission. The IMLS is the main 
provider of federal support for museums and libraries in the United 
States working with state and local organizations, its primary mission 
to strengthen the capacity for these institutions to link people, 
information, and ideas in a meaningful way. The fiscal year 2009 
LSTA grant offerings by the ISL were available in three categories: 
technology programs and services, resource sharing, and digital 
imaging projects. The digital imaging category suited the needs for the 
potential ICCA digitization project. 

The ISL's LSTA grant application and review process is 
competitive and requires the principal investigator to make the case 
for the value of the collection being considered for funding. Grant 
reviewers— experienced with digitization projects, past principal 
investigators, or project directors— evaluate proposals using a review 
rubric. The rubric is made available to grant writers and offers clear 
language about components of the grant application: project 
description, action plan, target audience, outcomes, timetable, 
personnel, and project sustainability (LSTA Grant Review Criteria, 
2010). The ISL provides further support to grant writers by hosting 
workshops designed specifically for digital imaging projects prior to 
the grant proposal deadline. The workshops cover planning the grant 
project, elements of the proposal, and the evaluation rubric. 

Developing the grant proposal afforded an opportunity to closely 
examine the ICCAs forty year history at ISU and to justify the 



388 From Argentina to Zambia 

significance of the collection in terms of how it related to historical 
and cultural trends on a state, national and international level. The 
resource, initiated by ISU's former University Museum Director, Dr. 
F. Louis Hoover, began with a collection of over 300 artworks 
gathered from the children of Illinois. Artworks from North American 
countries represent 55% of the current collection. Although ultimately 
international in scope, the seeds for the collection were sewn with 
creative works by the children of Illinois. With support from the 
University, the collection's scope grew under ISU art education 
professor Dr. Barry Moore (now emeritus), including artworks from 
children and adolescents of six continents, Antarctica being the only 
continent not represented. Relationship building and networking with 
international organizations such as the Christian Children's Fund 
reflected the University's longstanding record of supporting global 
engagement through activities including study abroad programs for 
students and welcoming international students to the ISU community. 

The LSTA grant application required detailed attention to 
planning the ICCA digital conversion including proposed timelines, 
personnel, and supporting resources. The project planning guidelines 
suggested through the proposal process largely reflect the detailed 
information for the digitization of cultural heritage materials 
thoroughly outlined in the Digitization Activities Project Planning and 
Management Outline document disseminated by the U.S. National 
Archives and Records Administration (Still Image Working Group, 
2009). 

Milner Library's Digitization Center, established in 2005 to 
primarily provide for the digitization needs of the University 
community, proved the logical choice for the digitization of the ICCA. 
The Center, equipped to digitize a variety of analog formats, provided 
contractual services for successful projects funded by LSTA grants 
awarded through the ISL: the Towanda History Project, a partnership 
between the Towanda District Library and Historical Society, and the 
Native American Collection, held at the McLean County Museum of 
History. These projects established the Center's reputation for 
partnering on LSTA grant funded digitization projects. Additionally, 
Milner Library and ISU's School of Art work collaboratively with the 



From Argentina to Zambia 389 



Center to produce an image database, ILSTUDIA, providing images 
for art, art history, design, and visual culture hosted on the library's 
CONTENTdm server. Further experience was gained with the 
digitization of two dimensional artworks when the Center digitized 
more than 250 prints from the New Editions Workshop project 
funded by the School of Art. 

The grant proposal emphasized the collection's value to provide 
primary source materials across disciplines including art, history, 
social sciences, education, English and psychology. The educational 
value of the collection was already established by the documented 
study and use of the resource by researchers (ICCA: Research, 2008). 
The digitized collection was not meant to replace the original 
artworks, but to facilitate access to a larger audience. Titled Imagine 
Illinois and Beyond: Celebrating Creativity Through the Eyes of Our 
Children! the grant proposal was submitted in May 2008 to the ISL's 
LSTA digital imaging grant program. In August the total amount 
requested, $85,934.00 was awarded for the project to digitize 8,600 
accessioned pieces in the ICCA. The grant cycle began October 1, 2008 
with all funds to be expended by June 30 th 2009. 

Constructing the Project 

The original grant proposal was submitted, approved, and awarded, 
allotting the majority of the funding for contractual services from the 
Digitization Center at Milner Library. Although the grant proposal was 
reviewed and approved at all required levels at ISU, the University's 
Grants Accounting and Comptroller's Office requested a budget 
amendment be filed after the grant was awarded, belatedly noting the 
Digitization Center named for contractual services could not be 
designated as a separate agency. An amendment to the grant budget 
needed to be filed to reflect expenditures to support the project. The 
budget amendment was submitted and approved by the University 
and the State Library. It was an unexpected hurdle to cross and 
created delays hiring grant funded personnel and ordering equipment. 
Additional complications were created by the reorganization of the of 
the Digitization Center, just weeks prior to the grant awards being 



390 From Argentina to Zambia 

announced. The director of the Digitization Center was reassigned to 
work full time on a separate external grant funded project, essentially 
leaving the ICCA project without a project director and the Center 
without a director. Without prior notice or planned options for time 
release from academic workload, the principal investigator took on the 
additional responsibility of project director for the digitization project. 
As author of the grant, the principal investigator had drafted the 
project planning that was already in place and the funded grant 
moved forward. 

Developing the Data 

When Milner Library began administering the ICCA, the resource 
came with a textual database holding museum generated information, 
as available, about the artworks including object identifiers, titles of 
the pieces, date created, artist gender, artist age, artist location, 
provenance, materials, dimensions and subject headings. The textual 
database laid the groundwork for collaboratively developing metadata 
schema, mapping, and standards used in the digital image database. 
Due to personnel turnover and position reassignments three different 
metadata librarians worked on the project during the planning phase 
of the grant, through the grant cycle, and after the grant cycle came to 
an end. The library supported filling the vacated positions so that 
metadata development continued to progress. 

Initially, guidelines for the LSTA grant required funded projects 
to upload their digital files and related metadata into the Illinois 
Digital Archives for dissemination. Milner Library was already using 
CONTENTdm to manage several other digital collections, and it is the 
choice of the Consortium of Academic and Research Libraries in 
Illinois. CONTENTdm has proven to be a reliable data storage 
solution for a variety of formats including images, documents, and 
audio. This is not a small consideration for an institution, not only due 
to cost, but also to preserving cultural heritage collections with 
concerns for moving the digital data attached to these resources 
reliably and robustly into the future (Van Den Bosch, Van Den Herik, 
& Doorenbosch, 2009). 



From Argentina to Zambia 391 



Vocabulary for the subject headings was most closely aligned with 
the Getty Art and Architecture Thesaurus (AAT). The project team 
continues to add and adjust subject headings as needed using the ATT 
as a guide, although not all terms are included in the thesaurus. For 
example, the terms "angels" and "ghosts" are excluded in AAT but are 
useful terms to apply in a children's art collection. Metadata for the 
collection was scrutinized closely after the images and information 
were matched and viewable simultaneously. Gaps in information were 
revealed. Descriptive metadata, such as titles for the artworks, were 
occasionally truncated due to the migration of data from one database 
format to another. This was an irregular occurrence which was 
remedied by consulting the original records or the artwork to 
complete the information. 

A related grant was written concurrently with the LSTA proposal 
requesting funding from the University's Research Grant program to 
support additional enhancement and development of descriptive 
metadata fields for the collection. For example, as the collection grew, 
accompanying artists' comments were documented on old key punch 
cards, sometimes typed, sometimes handwritten, in the 1970s and 
1980s by staff from the University Museum. The paper records, stored 
with the collection in the offsite library storage facility, provided no 
viable method of access for researchers. The awarded grant provided 
funds to hire a student who reviewed all paper records and 
transcribed more than 600 comments into a spreadsheet. Providing 
access to accompanying artists' commentary in addition to the images 
themselves provides researchers enhanced primary source material. 

When paired with the image of the artwork, the accompanying 
artist commentary enriches the descriptive metadata available to users 
and also potentially provides illuminating context for the image 
(Figure CHILD-i). The commentary from the young artists had not 
been made fully available in an accessible format to users in the past. 
This same grant also partially funded student assistants to transcribe 
and translate a collection of post World War II Germany artworks 
donated to the ICCA in the 1970s. 



392 From Argentina to Zambia 




Figure CHILD-1. To Live Underthe Sea, boy, 11 years old, Philippines, 1976. 
© International Collection of Child Art, Milner Library, Illinois State University, 
Normal, Illinois. Used with permission. Artist comments: "I wish to be friendly 
to all fishes under the sea so I could see the beauty and surroundings under 
the sea. I wish to stay forever under the sea. I will ride on the back of the 
shark. I will help the poor people and I will give them pearls." 

After the LSTA grant was fulfilled, a University Research Grant 
was awarded to support an ongoing project to write concise— one to 
two sentence— descriptions for the digital images from the ICCA. 
Descriptions were a missing piece in the information provided by the 
textual database. The accompanying descriptions will serve as an 
additional path to the end user, enhancing access and discovery. By 
Dublin Core definition, the description field provides a textual 
description for an image, such as a piece of children's artwork in this 
collection (Dublin Core, 2009). End-users, including students, 
educators, and researchers, will benefit from having additional access 
points available as supporting and valuable descriptive metadata for 
the digital collection of multicultural children's artworks. Although 
the majority of the 8,600 artworks converted under the LSTA 
digitization project have at least one term in the subject field for 



From Argentina to Zambia 393 



descriptive metadata, no information is available in the description 
field. The lack of information in this field limits the end-user's 
capacity to fully explore the resource for the purposes of teaching, 
learning, and research. 

The library's digital imaging specialists captured images of the 
artworks with the center's Betterlight Super 8K-HS camera working 
on a Tarsia 40 x 60 inch vacuum table which accommodated most of 
the pieces. Required specifications for scanning the artwork were 
provided by the State Library and were compatible with other digital 
projects the library had developed. The image files have an 
uncompressed TIFF master file (300 PPI), compressed JPEG image 
files (300 PPI), and a thumbnail GIF image (72 PPI). The digital 
imaging specialists and project student assistants resized and edited 
image files using Adobe Photoshop CS2. 

The Road More and Less Traveled 

The Digitization Center's location required the children's artwork to 
be transported from the university warehouse approximately three 
miles off campus to the library. Over the course of the grant cycle, 
approximately forty weeks, weekly trips were made to move artwork 
to the library and then return the artwork to the warehouse. 

Despite being located in the Midwest, climate never played a 
significant role in transporting the artwork, not a small consideration. 
A vehicle was able to drive into the warehouse for pickups and 
deposits to the library were delivered at a service door with a large 
overhang. The two dimensional works were transported in portfolios 
designed and crafted by the library's preservation staff. Each portfolio 
held 25-50 pieces depending on the size and material from which the 
artworks were made. 

In the last months of the grant cycle the project team began 
processing the traveling exhibits, about 500 artworks, framed and 
stored in shipping crates. These pieces required extra time and 
attention because many of the pieces had to be unframed, digitized, 
and then reframed. The exhibits posed more of a challenge for 
transporting back and forth because transportation had to be 



394 From Argentina to Zambia 

arranged in advance for University facilities staff to deliver the crates 
to the library and then return them to the warehouse. The process of 
transporting, uncrating, digitizing, reframing, general handling, and 
re-crating the artworks was more time consuming than anticipated. 

Library preservation staff worked with the project team from the 
planning phase and throughout the project cycle helping to identify 
workable solutions to related preservation and conservation issues. 
The overall condition of the artwork in the collection is fair, 
considering the age of the collection, some artworks more than fifty 
years old, and the materials used for children's artwork, typically not 
archival quality paper for example. A potential problem considered, 
during the planning phase, was the possibility that some of the 
artwork might have preservation or conservation concerns that would 
need attention. Working collaboratively with the library's preservation 
staff, the digitization staff handling the artwork created a rubric of 
potential preservation concerns and documented observations made 
about the condition of the artworks as each piece was handled that 
was unusual or in need of attention. As a result, the collection now has 
a record describing the condition of each piece that was marked for 
conservation attention. For example, several hundred of the artworks 
were in acidic mat board frames that necessitated removal as the 
pieces were handled. Excessive amounts of tape made this a more 
tedious, and time consuming, process than anticipated. Typical notes 
documented included: some water damage, tacky oil paint, fragile 
materials, acidic paper, tears, glue, tape etc. The digitization staff 
documented the preservation concerns throughout the process and 
consulted closely with preservation staff to make bridge fixes as 
needed. This documentation will be used to pursue grant funding for 
conservation and preservation of the artworks sometime in the future. 

Miscellaneous Malfunctions and 
Positive Project Outcomes 

Unanticipated technical problems can and will happen throughout a 
project period. Alternatively the positive outcomes provide an 
overpowering counterbalance. 



From Argentina to Zambia 395 



Scanning came to a halt for several days just one month into the 
LSTA project, when the Betterlight Super 8K-HS camera used to 
capture the artwork malfunctioned and had to be shipped to 
California for repair. The library's main server containing all of the 
project image files went completely down toward the end of the grant 
cycle resulting in six work days that proved difficult for processing, 
digitizing, and scanning the artwork. The project team adapted, saving 
files to the work station hard drives for temporary storage, until the 
servers were restored. After a change in library network configuration, 
the digitization center staff encountered an unexpected increase in the 
amount of time it was taking to open a TIFF file from the server. 
Moving image files between servers or from the server to a local hard 
drive was slowed significantly. 

Positive outcomes of the ICCA grant project and the increased 
level of digitization activity directly and indirectly influenced and 
contributed to changes in the Digitization Center. Repeated bending 
and lifting at the low level of the pre-fabricated scanning bed created 
ergonomic concerns. A frame to raise the scanner bed was designed 
and fabricated, raising the scanning bed to 30 inches thus improving 
the ergonomics of the setup and creating a more comfortable 
workflow. The library committed to supporting additional space by 
moving equipment and personnel from a cramped, shared area (which 
staff had uncomplainingly made functional) to a larger work space 
dedicated solely to the needs of the Digitization Center. The move also 
solved the slowed time to manipulate image files as the network ports 
and servers were upgraded from one hundred megabyte Ethernet to 
one gigabyte Ethernet in the renovated space. Staff increased as a 
second digital imaging specialist hired under the LSTA grant 
continued employment full time after the grant cycle ended. Two ISU 
students from the arts technology program were hired to work part 
time on the LSTA project. They assisted with scanning artwork, 
resizing images, and helped transport the artwork from the warehouse 
to the digitization center. 



396 From Argentina to Zambia 



Areas of Growth 

In 2007 Illinois Public Act 095-0307 altered the fabric of State 
awarded digitization grants. The newly enacted Illinois Information 
Technology Accessibility Act (IITAA) requires the preparation of 
audio descriptions to accompany digital images to enhance access for 
individuals with a print disability such as vision impairment. Due to 
the incongruous timeline of the fiscal year 2009 grant awards and the 
enactment of the IITAA, FY09 grant recipients, Milner Library 
included, were only asked to submit audio files to accompany twelve 
digital images. Funded applications henceforth are required to 
prepare audio descriptions for 100% of digital images. After attending 
an audio description (AD) workshop hosted by the Alliance Library 
System, the project director became intrigued by the concept of 
providing an alternate source of access to a visual resource. A small 
grant from the University was secured and provided seed money to 
hire a student assistant and begin researching the implications and 
mechanics of audio describing a digital image collection. Using the 
ICCA digital collection as a beta model, a pilot project has been 
initiated by the project director with the primary objective to 
investigate the challenging aspects of developing audio descriptions 
for an image collection. 

Audio description is defined as a "narrative technique that makes 
visual images more accessible to blind and low-vision people by 
producing audible written descriptions of non-verbal visual 
information" such as the digital images in the ICCA. The history of AD 
is rooted in the performing arts and dates back to the 1960s when a 
vision impaired employee from the United States Department of 
Education suggested preparing audio descriptions for films as well as 
the captioning already being provided for hearing impaired 
individuals. The idea was not supported however until the 1980s when 
advocates gathered and initiated an ongoing program sponsored by 
the Washington Ear, a group recording newspaper and magazine 
articles for the vision impaired and continued by providing 
descriptions for performing arts venues (Snyder, 2008). 



From Argentina to Zambia 397 



Employing basic principles for preparing AD suggested by the 
workshop, work began on writing transcripts for a selected number of 
images from the ICCA. The transcripts are brief, providing up to a one 
to two minute description averaging 200 words. Descriptive language, 
not analysis, attempts to provide a listener with a concise 
understanding of the image. The basic elements for the description 
use an introduction to the image, identify the primary theme or 
elements, and then describe in detail what is featured in the image 
(Figure CHILD-2). 




Figure CHILD-2. A Beautiful Afternoon with Birds, boy, 12 years old, 
Vietnam, 1972. © International Collection of Child Art, Milner Library, Illinois 
State University, Normal, Illinois. Used with permission. 



Audio Transcript: This artwork, a painting titled A Beautiful Afternoon with 
Birds Returning to Nests was created by a 12-year-old boy from Vietnam in 
1972. The centerpiece of this painting — composed primarily with blue hues — 



398 From Argentina to Zambia 

is two birds flying across a lightly speckled sapphire blue sky. A larger bird is 
painted in the upper left side of the artwork closely followed by a smaller and 
similarly painted bird. The beaks have only been suggested by a jagged line 
and the visible eye of each bird is a simple white dot. The avian bodies are 
sleek and graceful with plumage that subtly moves from light to deep 
turquoise. The birds blend closely with their inky blue-black shadows 
emphasizing their forked tail feathers, reminiscent of long fluttering coat-tails. 
This image is published by Milner Library at Illinois State University as part of 
the International Collection of Child Art Digital Collection. 

Approaches to uncovering resources and options for developing, 
writing, and recording narratives for images, grant funding, and 
working with units across the university are being explored. An option 
being initiated is building collaborative partnerships with teaching 
faculty to integrate the creation of audio description for digital images 
in the child art collection, into course curricula in related disciplines. 
The outcome of the pilot project research, still in the beginning stages, 
will inform future digital imaging initiatives integrating audio 
descriptions with digital images. The children's artwork presents a 
captivating invitation to construct an accessible and meaningful 
representation of an image with language. 

Conclusion 

When discussion about digitizing the ICCA began the motivating drive 
behind the idea was to provide improved access to support teaching, 
learning, and research. With approximately 97% of the collection 
digitized at the time of this writing, the ultimate objective of the 
project has been fulfilled, although outcomes have expanded and were 
influenced by a fluctuating climate of change that involved personnel, 
equipment, and content (ICCA Digital Collection, 2009). Because 
change is a constant in this environment a key component of a 
digitization project is a creative, cohesive, and flexible project team 
invested in the project. It is impossible to deny the overarching appeal 
this collection of children's art extends to all, including the staff and 
faculty across multiple library units who contributed to and supported 
the project. From a pedagogical perspective, the project provided an 
opportunity to engage numerous ISU students who contributed their 



From Argentina to Zambia 399 



time and talents to multiple aspects of the process including collection 
maintenance, digitization skills, translation of artworks, conservation 
work, and transcription of metadata. The students who worked with 
the collection were thrilled to have the experience of working with 
primary source materials and were quite excited to be a part of the 
process that brought the collection to digital life. In the academic 
environment, and especially at ISU, which prides itself at placing the 
learner at the heart of teaching and scholarship, the level of student 
engagement was an unanticipated and truly positive outcome. 

The project will continue with the ongoing refinement and 
enhancement of descriptive metadata, including expanded subject 
headings and description development. As the collection's identity 
shifts to the virtual, the responsibility to assure optimum access and 
searchability for continued use and research of this resource will 
remain at the forefront of its mission. Broadening the scope of access 
for a range of users by providing audio access will continue to be 
explored as the project moves into classrooms and provides 
collaborative opportunities for students to continue to contribute in 
meaningful ways to the project. The library will consider the forward 
direction other cultural heritage collections are moving in the virtual 
world (Ronchi, 2009). Embracing hypermedia to push the boundaries 
of the collection through virtual storytelling, gallery exhibits, and 
reconstructing the creative, historic, and cultural environment of a 
child artist are all possibilities. Ultimately this will serve the goal of 
remaining a meaningful and vibrant teaching and learning collection 
that will continue to inspire users researching the visual expressions 
of children. 

References 

Audio description Illinois. (2009). Retrieved March 13, 2010 from 
http://www.alsaudioillinois.net/ 

Best practices for creating digital files. (2010). Retrieved March 9, 
2010 from http://www.cyberdriveillinois.com/departments/ 
library/ what_we_do/ servicestechnologygrant.html 



400 From Argentina to Zambia 

Dragon, P. M. (2009). Name authority control in local digitization 
projects and the Eastern North Carolina postcard collection. 
Library Resources and Technical Services. 53 (3), 185-196. 

Dublin core metadata initiatives; Dublin Core metadata element set, 
version 1.0: Reference Description. (2009). Retrieved March 15, 
2010 from http://dublincore.org/documents/1998/09/dces/ 
Educating Illinois 2008-2014: core values. (2008) Retrieved March 
22, 2010 from http://www.educatingillinois.ilstu.edu/ 
plan_sections/core_values.shtml 

Hunter, N.C., Legg, K, & Oehlerts, B. (2010). Two librarians, an 
archivist, and 13, 000 images: Collaborating to build a digital 
collection. Library Quarterly, 8 (1), 81-103. 

Illinois digital archives guidelines for images. (2010) Retrieved 
March 14, 2010 from http://www.idaillinois.0rg/cdm4/guidelines/ 

Illinois general assembly. (2007). Illinois information technology 
accessibility act. Retrieved on March 14, 2010 from 
http://ilga.gov/legislation/publicacts/95/PDF/095-0307.pdf 

Illinois state library: Library services technology act review criteria 
and rubric, (n.d.). Retrieved on March 22, 2010, from 
http:/ /www.cyberdriveillinois.com/departments/library/what_we 
_do/pdfs/lstaio_rubric.pdf 

Institute of museum and library services: our mission, (n.d.). 
Retrieved on March 22, 2010, from 
http:/ /www.imls.gov/ about/ about.shtm 

International collection of child art digital collection. (2009). 
Retrieved on March 21, 2010, from 
http://tempest.lib.ilstu.edu/index_icca.php 

International collection of child art: research. (2008). Retrieved on 
March 20, 2010 from http://www.library.ilstu.edu/icca/research/ 

Lynch, C. (2007). Digital collections, digital libraries, and the 

digitization of cultural heritage information. In D.Kresh, (Ed.) The 
Whole Digital Library Handbook. Chicago: American Library 
Association. 



From Argentina to Zambia 401 



Stephenson, N. K. (2006). Preserving dissent: The Sam L. Slick 
collection of Latin American and Iberian posters. Art 
Documentation. 2 (1), 20-24. 

Van Den Bosch, A., Van Den Herik, J., & Doorenbosch, P. (2009) 
Digital Discoveries in Museums, Libraries, and Archives: Computer 
Science Meets Cultural Heritage. Interdisciplinary Science 
Reviews, 34, (2/3) 129-138. doi 10.1179/174327909X441063 

Ronchi, A.M. (2009). eCulture: Cultural content in the digital age. 
New York: Springer. 

Snyder, J. (2008). Audio description, the visual made verbal. In J. 
Cintas (Ed.), The didactics of audiovisual translation (192-197). 
Philadelphia, PA: John Benjamins Publishing Company. 

Still Image Working Group. (2009). Digitization activities project 
planning and management outline. U.S. National Archives and 
Records Administration. Retrieved March 12, 2010, from 
http://www.digitizationguidelines.gov/stillimages/documents/ 
Planning.html 

Acknowledgments 

Milner Library ICCA Digitization Project Team: Sara Caldwell, 
Digitization Center; Jim Caselton, Facilities; Anita Foster, 
Bibliographic Services; Ross Griffiths, Preservation; Erica Holden, 
Digitization Center; Christina Horna, Systems; Krena Hoyt, Systems; 
Leta Janssen, Administration; Heather Kosur, Preservation; Jason 
Paul, Systems; Patrice Andre Prud'homme, Bibliographic Services; 
Toni Tucker, Administration; Sharon Wetzel, Administration 

Milner Library student research and digitization project 
assistants: Daniel Abdalla, Stephanie Finch, Sheila Majumdar, Peter 
Nelson, Gina Pantone, R.J. Tortoriello 

Additional appreciations to Cheryl Asper Elzy, former Dean of 
University Libraries, Dr. Richard Satchwell, former Director, 
Digitization Center, and Alyce Scott, Digital Imaging Program 
Manager, Illinois State Library. 



Special Collections, Digitization, and the 
Classroom: A New Model 

Mark Phillipson (Columbia Center for New 
Media Teaching and Learning) 

Michael Ryan (Columbia University Rare 
Books and Manuscripts Library) 



Abstract 

The Black Radical Archive is a small pilot project at Columbia 
University that leveraged digitization to involve students in the 
discovery and description of heretofore hidden collections 
(http://blackradicalarchive.ccnmtl.columbia.edu/). This project is the 
result of innovative collaboration between archivists, educational 
technologists, a faculty member deeply engaged with archival 
collections, and his students— who made selections of their own and 
added to the digital archive. In this chapter we describe cross- 
divisional support for the project, its implementation in a Spring 2009 
seminar entitled "Black Radicalism and the Archive," and lessons 
learned from the informal, just-in-time digitization intrinsic to course- 
driven repository building. 

Keywords: Activism, Activists, Archives, Customization, Digital 
repository, Teaching, Drupal, Hidden collections, Informal 
digitization, Learning, Multimedia, Pedagogy, Processing, Repository, 
Special collections, Support, User-contributed. 



Introduction 

Special collections in many ways define the character of an academic 
library. They are rich and sometimes undiscovered islands of unique 



402 



Special Collections, Digitization, and the Classroom 403 

materials amid an ocean of more generally available information 
resources. 

Locally held archival materials help define the identity and 
character of their parent institutions, attracting and nurturing 
research affiliations and communities. Faculty members who have 
actively benefited from direct access to such materials in their own 
research are looking for practical ways to involve their students in the 
experience and excitement of working directly with primary source 
materials. 

In this chapter, we discuss a pilot project at Columbia University, 
the Black Radical Archive (2009), that sought to integrate special 
collections into the classroom — redefining some traditional notions 
of the academic archive and its use. This project supported a Spring 
2009 graduate seminar's focus on three special collections held by the 
Columbia University's Rare Books and Manuscript Library (RBML). 1 
It facilitated a deeper level of materials-based discovery and research, 
providing students the means to discover, categorize, annotate, 
digitize, and share important holdings in the collections. In the 
process, the visibility of the holdings and archival practice was raised, 
along with further prospects for drawing on archives to support 
teaching and learning. 

Though digital surrogates are sometimes faulted as a replacement 
for tactile contact with original material, the Black Radical Archive in 
fact leveraged digitization to encourage hands-on engagement. At the 
height of the project's implementation, the Black Radical Archive 
shifted in function from a repository of digitized items to a hub for 
uploads of additional items discovered and digitized by students. The 
project, then, emphasized the importance of physically inspecting 
unique materials, even as it cultivated the advantages that digital 
surrogates offer for convenient and repeated inspection, communal 
access, and non-invasive annotation. 



1 RBML's holdings and activities may be seen online at 
http : / / www. columbia.edu / cu /lweb / indiv/rbml / . 



404 Special Collections, Digitization, and the Classroom 

Exposing the Collections 

When he conceived of a graduate-level seminar exploring the 
intersections of activism and archive called "Black Radicalism and the 
Archive," Brent Edwards, Professor of English and Comparative 
Literature at Columbia, tailored it to three collections held by RBML: 
the Hubert H. Harrison Papers, a fully cataloged collection with a 
published finding aid; the C.L.R. James Papers, a collection of 
heretofore disparate collections that was in the process of being 
assembled as the seminar met; and the Amiri Baraka Papers, a large 
unprocessed collection with only a preliminary carton survey. Despite 
these different levels of cataloging and indexing, Edwards wanted his 
students to discover, categorize, annotate, and compare material from 
all three collections— developing primary materials-based research 
skills in the process. 

Edwards's course occurred at an opportune time for both RBML 
and the students in the seminar. Special collections librarianship has 
changed markedly in its emphases during the past decade or more. 
Without losing its curatorial, collections-based focus, the field has 
invested heavily in promoting outreach and in directly supporting the 
work of teaching and learning on campuses. 2 Special collections units 
are now more open to a diverse constituency than they used to be, 
more concerned with creating a broad base of users. Outreach 
librarians are now common in special collections units, as are class 
and seminar rooms. As library staff collaborate more with faculty on 
course design, primary source materials are increasingly integrated 
into undergraduate and graduate courses. At Columbia and many 
other institutions, curators actively reach out to faculty for the 
purposes of better integrating source materials into curricula; RBML 
hosts two or three classes per day during the academic term, each of 
which use rare books, manuscripts, documents, and other materials 
from the collections. 



2 For background on these efforts, see Traister, D. (2003), Smith, S. 
(2006), and Association of Research Libraries (2008). 



Special Collections, Digitization, and the Classroom 405 

As a result of such outreach efforts, archival materials are better 
integrated into courses as enrichment (library staff presentations of 
source materials being studied in a course in later editions and 
formats), research consultation (in connection with assignments 
requiring students to consult books or manuscript materials in special 
collections), or surrogate access (digital facsimiles of selected special 
collections material made available to students through course 
management systems, or CMSs). When Professor Edwards 
approached RBML about his Black Radicalism seminar, we assumed 
that he would be interested in the classroom support that RBML is 
now accustomed to providing: that is, a review of sources pertinent to 
the course, a display of some of them, and some arrangement for 
ongoing access to select materials in the RBML reading room or the 
course CMS. We were wrong. 

Professor Edwards was thinking outside the box — or, more 
accurately, he was thinking deep inside the box, the archival storage 
unit. Edwards was interested in providing students with access to 
primary source materials before they were organized for use. He 
wanted his students to confront and try to make sense of historical 
objects that had not been rearranged by archivists. Of course it is 
common practice in research libraries not to open collections for use 
until they have been fully processed. But again, Professor Edwards 
approached RBML at an opportune time. Like other large research 
libraries, Columbia has its own formidable backlog of unprocessed 
and underprocessed archival and manuscript collections. Addressing 
these arrearages became an ARL priority in the late 1990s, and the 
subsequent momentum to provide new energy and funding streams 
resulted in the "Hidden Collections" initiative that continues to play a 
vital role in our research libraries. 3 Echoing ARL recommendations to 
"connect the exposure of hidden collections to ARL's strategic priority 
for Research, Teaching, and Learning," (Special Collections Task Force 



3 See, for example, Greene and Meissner (2005), and Association of 
Research Libraries: Exposing Hidden Collections (2009). 



406 Special Collections, Digitization, and the Classroom 

Final Report, 2006), Professor Edwards asked us to challenge a long- 
standing policy of processing a collection before opening it up for use. 

Edwards, of course, was less interested in library policy than in 
pedagogical opportunity. His syllabus framed variously (un)processed 
materials as a chance for students in the seminar to make original 
discoveries and contribute to a scholarly effort to describe the 
collections: "Part of the unique challenge — and, hopefully, the 
excitement — of this seminar is that the three collections we will be 
working with are in different states of organization. At times, 
especially with the Baraka Papers, we will be investigating boxes with 
little or no information about what we might find in them. On the one 
hand, this means that to a certain extent we will have the opportunity 
to discuss the collecting practices (inchoate as they might sometimes 
be) of these intellectuals themselves, as they gathered and stored a 
range of materials over many years. On the other hand, this means 
that, by noting the layers of material in a given box, by tracking 
sources, by deciphering handwriting, by dating an artifact or 
manuscript, by annotating and explicating, in the long run we will be 
helping RBML in the effort to catalog and make available to future 
researchers these enormous and currently unwieldy collections. " 

For Edwards's purposes, each collection was ideal for the sort of 
practical engagement and various theoretical issues attending the 
formation of archives that he had in mind. The C. L R. James Papers 
were really an artificially constructed assemblage of materials by and 
relating to James that had been harvested from a variety of sources in 
North America and the UK. One of the most important social and 
political philosophers and activists in the 20 th century, James led a 
peripatetic life. He was careless about his personal effects, choosing 
instead to put his energies into philosophy, literature, and global 
politics. Friends and acolytes salvaged what they could along the way, 
shaping his legacy in this way, and those efforts resulted in what 
forms the core of the James Papers at Columbia. At the other extreme, 
Amiri Baraka has lived in the same house in Newark NJ for more than 
fifty years. He is also an inveterate collector, harvester, and saver. But 
he is not a disciplined or organized one, and so his papers, while 



Special Collections, Digitization, and the Classroom 407 

voluminous, reflect an idiosyncratic self-archiving that could be called 
comprehensive, organic, or simply disorganized. 

Black Radicalism and the Archive, then, would be an ambitious 
and wide-ranging course. It would focus students on the material 
aspects of archives that document activist movements, as well as on a 
larger set of theoretical concerns entailing the formation of political 
and cultural identities. Working with the class on this level was a new 
and exhilarating task; the challenge was to derive from material 
archives the making of a broader movement. As Edwards planned his 
course and conceived of readings and assignments trained on these 
materials, it became clear that the collective research he had in mind 
would require its own unique and equally ambitious support. At the 
center of that support would be a digital workspace, allowing him and 
his students to select, organize, share, and appraise each other's 
findings from three very different collections. 

A New Model of Support 

Once Professor Edwards conceived of a virtual space in which to 
coordinate and analyze the range of materials that students in his 
seminar would be considering, he turned to another division of 
Columbia University Libraries' Information Services: the Columbia 
Center for New Media Teaching and Learning (CCNMTL). CCNMTL's 
mandate is to facilitate and advance teaching at Columbia through the 
purposeful use of new media. Over its ten years of operation, 
CCNMTL has grown into a service enterprise that supports over 4,000 
instructors at Columbia University. 4 It was the natural group to help 
design and implement what would become termed the Black Radical 
Archive, a digital workspace for Edwards's seminar. 

Serendipitously, CCNMTL's priorities had been evolving 
congruently to RBML's; each group was exploring new ways to 
incorporate collections into teaching and learning. In 2007 CCNMTL 



4 More information about CCNMTL's services and projects is online at 
http://ccnmtl.columbia.edu. 



408 Special Collections, Digitization, and the Classroom 

had launched a strategic initiative called Digital Bridges 5 specifically 
devoted to the development of innovative connections between 
curated collections and classroom-based study. Digital Bridges 
projects have resulted in a variety of educational environments and 
tools that draw on collections held by museums, public media 
producers, documentary filmmakers, scientific laboratories, and 
academic libraries. As it cultivates new ways to incorporate collections 
into teaching and learning, CCNMTL is increasingly reliant on 
collaborative relationships formed with Columbia University Libraries 
units such as RBML. 

Supporting Edwards's course, in fact, drove new levels of 
coordination between RBML and CCNMTL. Though each division 
digitizes materials as part of its service mandate, the purposes, time- 
frames, and quality levels of such digitization vary to a great extent. 
Digitization of source materials is, of course, intrinsic to CCNMTL's 
work supporting multimedia instruction at Columbia, and yet much of 
this digitization is conducted 'just in time' within the context of an 
individual semester or project; it does not conform to archival and 
preservation standards. At the same time, RBML conducts a robust 
digital program of its own. Partnering with Columbia's Libraries 
Digital Program Department (LDPD) and the Preservation and 
Reformatting Department (PRD), RBML's digitization is performed to 
high quality standards and often tied to extracurricular exhibitions 
and events. Setting up and stocking the Black Radical Archive, then, 
entailed coordinating these various digitization practices across 
CCNMTL, RBML, and units in the library supporting preservation- 
level digitization. 

Shortly before the semester met, Edwards worked with RBML 
staff to identify items that he wanted to make available by dates 
pegged to his syllabus. The standard workflow for digital exhibition of 
RBML materials involves selection by curators, digitization to 



s Background on Digital Bridges and description of various projects 
developed by this initiative is available at http://ccnmtl.columbia.edu/ 
digitalbridges. 



Special Collections, Digitization, and the Classroom 409 



preservation standards by PRD, and design and presentation of 
derivative versions of the content by LDPD. But in this case, the 
tighter deadline was more typical of CCNMTL's course-specific 
digitization or an e-reserves process than a standard high-quality 
digitization workflow; Edwards needed a turnaround of materials 
scanned and posted to the digital archive in as soon as one week, in 
some instances. This compressed timetable was partly due to the 
unprocessed condition of some of the items: it took Edwards and 
RMBL staff some time to conduct an informal survey of uncataloged 
materials that would be pertinent to the course. 

When there was enough lead time, Edwards's initial selections 
from the collections were processed in the traditional way: digitized in 
the libraries' preservation lab to archival standards, with a derivative 
version passed on to CCNMTL to post to the Black Radical Archive. In 
many cases, though, the pressing deadlines of Edwards's syllabus 
meant that shortcuts and workarounds were necessary. Edwards's 
engagement was crucial here: whenever necessary, he was ready to 
spend time in the collections himself, taking digital pictures for the 
Black Radical Archive. He was thus able to give his class online access 
to items from collections whose traditional digitization would have 
taken a long time - if it were even attempted. For example, Edwards 
was interested in a series of scrapbooks assembled by Hubert 
Harrison; these scrapbooks helped underline the course's emphasis on 
this activist's quirky and suggestive arrangement of materials. 
However well-suited such scrapbooks were for the course, archival- 
level digitization of such complex objects is a daunting, labor- 
intensive prospect, necessitating correlation of page components and 
layers. Assisted by RBML curators, Edwards conducted his own, 
decidedly non-archival imaging of these scrapbooks. 

While this survey was taking place, CCNMTL was also consulting 
with Edwards to set up the Black Radical Archive, the website in 
which digital surrogates generated for- and eventually by- the class 
would collect. As a digital hub supporting student investigation of 
these collections, the Black Radical Archive was structured to help 
students quickly find as well as contribute items. Selecting an easily 
modifiable platform, CCNMTL built the archive using a lightly 



410 Special Collections, Digitization, and the Classroom 

skinned instance of the open source Drupal content management 
system. Drupal allowed for quick set-up and modification of content 
types, as well as suitable administrative access for the fifteen or so 
students enrolled in the course, some of whom were visiting from 
other institutions. 

The website was set up and integrated with Columbia University's 
authentication system shortly before the semester began. An alternate 
form of registering was also appended to the site in order to 
accommodate the visiting students. Authenticated students were 
allowed to search and access all content posted in the archive, post 
comments, browse and edit metadata, and upload files. A higher level 
of administration was reserved for CCNMTL so that the archive could 
be further designed, modules updated, and user roles managed. The 
interface was set up so that students could filter records by collection, 
scan all collection locations, and quickly browse assets by column 
sorting of titles and media types. 

To support locating and indexing items from the collections, and 
the box descriptions that would be the seminar's capstone assignment, 
CCNMTL agreed to structure the repository with a simple and 
customized data model. In this model, every asset uploaded into the 
archive had to be connected to a location. Usually these locations were 
boxes (with a numbering convention defined for the class). Because 
assets from outside the three collections might also be uploaded to the 
digital archive, a generic "Outside" location category was also created. 
Metadata for the assets was kept as simple as possible — far simpler 
than typical library description — to help make the eventual student 
input of new assets into the archives as frictionless as possible. 

Shortly before the class met, CCNMTL worked closely with 
Edwards to stock the archive with the many pictures he had shot of 
scrapbooks and other material he had photographed in the Hubert 
Harrison collection. Since some of these files were too large to be 
directly uploaded into the Drupal, CCNMTL arranged supplementary 
server space through the Libraries Digital Program (LDP) at 
Columbia. This too was a first: LDP had never before stored files that 
were not generated by professional staff at preservation-level 



Special Collections, Digitization, and the Classroom 411 

standards. 6 Since Edward's selections for the digital archive included 
aging audio and video tapes of various formats, several other 
variations on traditional digitization processes occurred during the 
semester as RBML and CCNMTL staff continued to add assets to the 
digital archive, using Edwards's syllabus as a roadmap. RBML sent old 
audio tapes out to a third-party vendor; CCNMTL, more experienced 
with video processing, digitized some VHS tapes and, in the instance 
of some badly damaged tapes, also commissioned an outside vendor. 
All of these assets, generated through various means, were uploaded 
into the Black Radical Archive in time for students to access for 
specific class sessions. The Black Radical Archive, then, grew as a 
result of diverse support interactions and inputs, ranging from 
archivists to the instructor to various preservation specialists to 
educational technologists and, finally, the students in this unique 
course. 

Classroom Support and Implementation 

Class sessions were all held in a conference room in RBML, and 
supporting these sessions was a logistical challenge in its own right — 
involving coordination of cartons of material, presentations by 
students of their discoveries, and collective access to the Archive 
during class discussions. Participation of RBML staff in the class was 
critical. Since almost none of the students in the course had any 
significant experience with or exposure to primary source materials 
prior to the course, an RBML archivist who was assigned to 
supporting the course gave the group as a whole a short primer in 
archives and archival management, reviewing with them the 
fundamentals of acquiring, surveying, processing, and providing 
access to archival collections. In addition, she met with students 
individually throughout the course to answer their questions and help 



6 In moving in-house library digitization support towards more informal 
ground, this project resonated with calls from Davidson (2009), Dooley 
(2009), and many others to prioritize services for an expanding user base for 
special collections; Dooley urges us to "digitize with abandon." 



412 Special Collections, Digitization, and the Classroom 

them with their assignments. In effect, she served as a TA without 
portfolio. This as-needed support was provided in lieu of more formal 
training of Edwards's students in archiving practice, which would 
have overloaded a syllabus already rich with cultural theory and 
historical background about the seminal activists being studied. The 
goal of the course, after all, was to get students engaged with archival 
practice as a point of entry into more theoretical study of the practices 
and influences of Harrison, James, and Baraka. 

Another component of class support was the accommodation of 
important guest speakers. In addition to noted scholars and experts, 
Baraka himself visited one session and explained his own "archival 
techniques," thus providing a further layer of documentation and 
information to add to the Black Radical Archive. Each of the guest 
sessions was taped, and the tapes became part of the evolving 
narrative of activism and archiving constructed on the foundations of 
the actual primary source materials, affording further material for the 
growing digital archive accompanying Edwards's course. 

In the second half of the semester, Professor Edwards gave the 
class instruction on uploading items into Black Radical Archive. 
Students were then ready to undertake the culminating assignment: 
each one was assigned a box in the unprocessed Amiri Baraka 
collection and then asked to inventory the contents of their box. To do 
so, they would consult with RBML archivists, take digital pictures of 
particularly interesting items they discovered in their box, and upload 
box inventories and item photographs into the digital repository. 
Though Edwards felt that this assignment could yield some interesting 
information for RBML, in the form of preliminary inventories of 
uncatalogued boxes that were modeled on inventories of processed 
collections, the goal was not to produce authoritative documentation. 
Instead, students were asked to produce lists of contents, with as 
much contextual information (such as the date of an event, or the 
identity of a figure in a photograph) as they could ascertain. These 
unofficial selections and annotations by students joined the more 
professionally catalogued and digitized material already uploaded into 
the Black Radical Archive. 



Special Collections, Digitization, and the Classroom 413 

Even though Edwards's seminar met regularly in a conference 
room in RBML and could retrieve and display physical collections, 
digital surrogates from the archive were often displayed instead of 
actual objects during class. This was a matter of convenience and 
efficiency; discussion could engage details from many objects and 
could proceed in spontaneous directions without being bogged down 
by physical retrieval and replacement of items. Some items were 
simply too fragile and cumbersome for the students to pass around; 
the digital archive facilitated quick pinpointing of specific parts of 
these objects during discussion, without loss of the visual complexity 
that would have been hard to perceive in, say, photocopies. 
Digitization of a variety of old audio and video formats meant that the 
class could collectively consider such multimedia items without 
having at hand various playback instruments. The digital archive also 
facilitated convenient distribution of long manuscripts in the 
collections, which were typically printed out by individual students for 
reference before class and brought in for group perusal. 

Students later described the "contagious" energy of their own 
uploading to the Black Radical Archive of discovered items from the 
Baraka boxes: their digital staging of material that would then be 
contextualized by findings in other boxes. This staging spurred 
discussions about theoretical and practical considerations of 
archiving. Items with particularly complex provenance gave rise to 
particularly lively conversation during the seminar. For example, the 
Baraka boxes included audio tapes with liner notes scrawled on them 
by the likes of Nina Simone and Allen Ginsburg— tapes that may very 
well have been owned by these other luminaries at some point. 

Collective appraisal of such objects, facilitated by the Black 
Radical Archive, helped students think about the complexities of 
provenance, the layers of interventions in any archive by original 
collectors and subsequent organizers of material, and the artificiality 
of any one preserved state of a collection. One student ambitiously 
documented each "layer" that he uncovered in his box as he dug 
through it rather like an archaeologist, and shared these layers with 
the rest of the class to spur discussion of patterns of proximity across 
the Baraka collection— patterns that a classmate went so far as to term 



414 Special Collections, Digitization, and the Classroom 

the "poetics of the archive." More simply, in-class consideration of 
items uploaded by students helped them compare contents of their 
assigned box to those in other boxes, and therefore gain various 
chronological, thematic, and material views of the unprocessed 
collection. 

Participation in the course not only immersed Edwards's 
graduate students in the practice of archival research, it also suggested 
to them the usefulness of digital surrogates as a component of such 
scholarly practice. One student discussed the way that a preliminary 
inventory of items scanned by the instructor and classmates helped 
him to identify patterns and target investigation of specific objects in 
the archive — in short, to make more efficient use of his time with the 
physical collections. Another student was prompted to think about the 
benefits of informal, collaborative digitization for research in a subject 
area she is researching, modern Egypt; items of interest for this topic 
are as likely to be discovered on the likes of eBay as in a formal 
archival collection, and a digital hub, this student speculated after 
working with the Black Radical Archive, could help scholars 
distributed around the world coordinate findings. 

Though in-class use of the Black Radical Archive was heavy, very 
few annotations of individual objects were posted on the site, despite 
the site's capacity to connect annotations to uploaded items. This was 
partly due to the structure of assignments in the class; by the time 
students were engaged in digitizing items from the unprocessed 
Baraka collection and comparing the contents of their assigned boxes, 
the semester was drawing to a close; annotation of various objects 
happened verbally as a matter of class discussion. In addition to a 
main projector and screen, several individual laptops anchored 
discussion with reference to various items in the digital archive. Class 
discussion anchored theoretical readings to the items stocked by 
Edwards in the Black Radical Archive and, as the semester proceeded, 
it was increasingly concerned with targeted identification of 
similarities and patterns across the collections. Student annotations 
that were uploaded into the archive, then, tended to be appraisals of 
groups of objects rather than of individual objects themselves. 



Special Collections, Digitization, and the Classroom 415 

As one of the few student who came into the course with previous 
experience with archives stated, the digital archive served as a 
"collective interpretive project," not a "collective digitization project." 
Had the class's emphasis been on building an online library, rather 
than on sharing an arena of interpretation, students would have 
concentrated on assembling and posting deeper metadata. Certainly 
the metadata in the Black Radical Archive is at present too sparse to 
serve researchers outside the course; it is limited to item titles, 
collection and box locations, and supplementary notes. Asking 
students in a future class to enrich the digital archive with more 
rigorous and informative descriptive metadata could be an interesting 
further educational use of the archive, should Edwards wish to further 
develop it as a resource for other researchers, or it could be a task 
assigned to a graduate student working with RBML archivists. 
Edwards's students were intrigued by the idea that their selections 
could help inform subsequent work with the three Columbia 
collections, though they agreed that subsequent cataloging would be 
necessary to make the digital archive useful to researchers beyond 
their class. Their most substantial legacy posted to the Black Radical 
Archive was, in fact, item-level description of unprocessed boxes, 
descriptions that ran to notable length in some cases and proved 
detailed enough to assist RBML in its eventual processing of the 
Baraka Papers. 

The Archive beyond the Class 

The Black Radical Archive survives the course it was built for, not as a 
comprehensive representation of the collections it draws on, but, 
rather, as an expanding documentation of engagements with these 
archives. It will thus raise awareness and interest in these rich 
collections, and, we hope, entice more teachers, students, and 
researchers into direct contact with archival holdings. 

From the perspective of a special collections unit nested within a 
research institution, the lessons learned were several and significant. 
First, it challenged two major sets of operating policies and practices: 
to wit, that unprocessed collections should not be made available prior 



416 Special Collections, Digitization, and the Classroom 

to processing, and that digitization projects should only by designed 
and undertaken for fully processed materials with stable metadata. 
Opening up unprocessed collections to students proved to be 
pedagogically and archivally important. In giving students access to 
primary sources prior to library mediation, the class allowed students 
to participate in the literal making of the historical record. At the same 
time, it generated descriptive information at a highly granular level 
that will eventually be of great use to processing archivists. The 
detailed box content lists provided by the students proved to be of a 
high quality; given the fact that the students brought to their tasks a 
depth of subject knowledge not typically available to a processing 
archivist, their contributions to the future organization of these 
collections will be important. Although it will not be possible to do this 
across the board in the future, involving students and scholars in the 
description of collections prior to their processing is an option we 
should try to exploit on a more regular basis — whether or not it is tied 
to an actual course. 

Digitization projects in the Columbia Libraries are complex, 
formal, and labor-intensive. Standardization across the system is 
rigorous and necessary if the products are to be integrated into a 
larger common pool of digital assets. "Scan once not twice" is dogma. 
Such projects are thus developed in a team environment, subjected to 
a demanding vetting process, and implemented in a highly controlled 
fashion. These projects typically reflect some consensus within the 
Libraries as to which materials would be of most use to researchers 
when presented on the web. The assumption is that if we build it, the 
research community will find it and use it. 

The lesson we learned from the collaboration is that we need to 
accommodate a more diverse set of needs for the generation of and 
access to digitized material. Aside from the benefit of surfacing 
previously hidden holdings, there may be good pedagogical reasons 
for doing so. Digitizing for the Black Radical Archive was entirely 
user-defined; RBML staff had no role in the selection process other 
than facilitating access and offering advise in response to the course 
instructor and students. The scans that resulted were of considerable 
value to the course, and they will always be available in the future 



Special Collections, Digitization, and the Classroom 417 



should they need to be repurposed in some way. RBML has worked 
with CCNMTL on many occasions in the past in contributing to web- 
based course support. What was new here was that the content of that 
support was driven by users, not librarians. This project was bottom- 
up, and its success encourages us to look for future opportunities to 
re- engineer RBML's workflow to dovetail with CCNMTL' s digital 
facilitation of learning activities. 

This type of project is especially staff- intensive and time- 
consuming. Archivists assigned to the course had to invest 
considerable time working with Edwards on selecting materials, with 
other staff helping to process digitization, and with students in 
Edwards's course using these collections on a weekly bases. Moreover, 
they needed to attend every session of the course. CCNMTL staff had 
to be in constant communication with Edwards, RBML, other library 
units, and outside vendors to keep the digital archive coordinated to 
an unfolding syllabus. In the age of "sustainability" and "less process, 
more product," this customized support sins in many ways. But while 
it is not a working model for day-in, day-out course support, it does 
provide interesting options for boutique projects, and it serves to 
remind us how badly we need the boutique along with the mainstream 
and the generic. Research institutions are congeries of expertise, and 
special collections units need actively engage with complementary 
types of expertise —such as that cultivated by instructional technology 
groups, as well as the faculty and students we all serve— as part of 
their routine work. 

Developing any course ab ovo is arduous and challenging; as 
anyone who has ever taught knows. But now that it has been 
developed and a model of it remains via the Black Radical Archive, the 
course — and intradivisional relationships formed in support of it — 
can be reused in various ways and for various purposes. We look 
forward to trying similar approaches with other faculty using a 
different menu of materials for different purposes. In the meantime, 
informal digitization and annotation of unprocessed material 
contributed to the forming of a unique engagement with our 
collections, and our task going forward is to preserve and make 



418 Special Collections, Digitization, and the Classroom 



accessible this effort of interpretation as part of the Harrison, James, 
and Baraka papers. 

Special collections and educational technology units considering 
these types of partnerships should be prepared to devote significant 
staff resources to them. In our view, the effort is well worth it, since it 
directly integrates unique holdings, faculty and students, and the 
diverse capacities of support units into the heart of our institution's 
academic mission. And that, after all, is why we are here. 

The authors would like to thank Brent Edwards, Susan Hamson, 
Laura Helton, Lea Osborne, Patricia Renfro, and Lytton Smith for 
their assistance with this article. 

References 

Association of Research Libraries. (2006). Special collections task 
force final status report, 2006. Retrieved March 29 from 
http://www.arl.org/rtl/speccoll/spcolltf/statuso706.shtml 

Association of Research Libraries. (2006). SP296 Public services in 
special collections, November 2006. Retrieved March 29 from 
http://www.arl.org/bm-doc/spec296web.pdf 

Davision, S. (2009). If we build it, will they come? Strategies for 
teaching and research with digital special collections. RBM, JO(i), 
37-49- 

Dooley, J. (2009). Ten Commandments for special collections 
librarians in the digital age. RBM, 10 (1), 51-9. 

Greene, M. And Messner, D. (2005). More product, less process: 
Revamping traditional archival processing. American Archivist, 
68(2), 208-263. 

Prochaska, A. (2009). Digital special collections: The big picture. 
RBM, 10(1), 13-24. 

Smith, S. (2006). From "treasure room" to "school room": Special 
collections and education. RBM, 7(1), 31-9. 



Special Collections, Digitization, and the Classroom 419 



Traister, D. (2003). Public services and outreach in rare book, 
manuscript, and special collections libraries. Library Trends, 
52(1), 87-108. 



Part IV - One Plus One is 
Greater Than Two: 

Collaborative Projects 



Digital Treasures: The Evolution of a 
Digital Repository in Massachusetts 



Dodie Gaudet (Central MA Regional Library System) 
Kristi Chadwick (Central/Western MA Automated 

Resource Sharing System) 
Jan Resnick (Western MA Regional Library System) 



Abstract 

Digital Treasures is a digital library collection of the history of central 
and western Massachusetts. It is a collaborative project among 
Central/Western MA Automated Resource Sharing System (C/W 
MARS), Central MA Regional Library System (CMRLS) and Western 
MA Regional Library System (WMRLS). Initiated by C/W MARS in 
2006, Digital Treasures began as a pilot program when C/W MARS 
purchased equipment and software and set up a scanning lab at its 
headquarters in Worcester. Currently Digital Treasures has 36 
collections from libraries, with over 1,300 accessible images. C/W 
MARS, CMRLS and WMRLS continue to collaborate on ways to bring 
funding, selection guidance and metadata expertise to their member 
libraries and bring access to the wealth of cultural history of the 
Commonwealth. 

Keywords: Central Massachusetts Regional Library System 
(CMRLS), Central/ Western Massachusetts Automated Resource 
Sharing (C/W MARS), Digital Barn Raising, Digital Treasures, Digital 
Commonwealth, Western Massachusetts Regional Library System 
(WMRLS) 



423 



424 The Evolution of a Digital Repository in Massachusetts 



History & Background 

In the early 1960s, three Regional Library Systems were formed in 
Massachusetts to provide a variety of services directly to the public 
libraries in the western, central, and eastern parts of the state. The 
services were (and still are) provided at no cost since the Regions are 
funded by the state. When automation began entering libraries some 
20 years later, the WMRLS (Western Massachusetts Regional Library 
System, 2010) and CMRLS (Central Massachusetts Regional Library 
System, n.d.) were instrumental in helping to establish a fee-for- 
service automated network available to all types of libraries named 
Central/ Western Massachusetts Automated Resource Sharing or C/W 
MARS (C/W MARS Inc, 2009). 

In 1998, the Regional Library Systems became multi-type. All 
public libraries are automatically members of a Massachusetts 
Regional Library System (MRLS). All academic, school, and special 
libraries are eligible for membership if they meet basic qualifications. 
Any WMRLS or CMRLS member can join C/W MARS by paying the 
membership fees. Not all libraries choose to join C/W MARS: some 
are small and cannot afford the fees; others are larger, have their own 
network staff and prefer a stand-alone Integrated Library System. The 
number of libraries in WMRLS and CMRLS is larger than the number 
in C/W MARS. As of March 1, 2010 membership in the three 
organizations broke down as follows 





CMRLS 


WMRLS 


C/W MARS 


Academics 


17 


19 


11 


Publics 


72 


103 


130 


Schools 


120 


164 


2 


Specials 


26 


25 


4 


Total 


235 


311 


147 



With the imminent development of Digital Commonwealth, a 
state-wide portal and digital repository, CMRLS and C/W MARS were 
both interested in setting up a small lab so that libraries could begin 
scanning photos, documents and other items which could also be 



The Evolution of a Digital Repository in Massachusetts 425 

discovered via the state-wide portal. A Regional Library System and 
an automated network in the eastern part of the state had each 
purchased a scanner and offered instruction on its use, but few 
libraries took advantage of the opportunity. Considering this, C/W 
MARS proposed that one of its tech-sawy staff members who had an 
interest in digital photography be in charge of the lab and do the 
actual scanning if local library staff supplied the metadata. CMRLS 
had a cataloger on staff who could conduct workshops on Dublin Core 
and be available for metadata consultation. Since the offices of the two 
organizations are a mere seven miles apart, joining forces to support a 
single repository made the most sense and Digital Treasures (see 
Digital Treasure, 2009) was born. 

The Scanning Lab 

Due to some new memberships and some existing members 
upgrading, C/W MARS had enough discretionary funds to purchase 
hardware and software and create a new part-time position. The 
Executive Hardware for the new lab consisted of an Epson Expression 
10000XL flatbed scanner with an 11" x 17" bed. A new server had 
recently been purchased for the network's regular activity and the old 
server was used for Digital Treasures. After much research, OCLC's 
CONTENTdm was chosen for the creation, storage and access to 
digital materials. Anticipating the future growth of Digital Treasures 
and the inclusion of text, OCR extensions were purchased at that time. 
Other programs loaded to the digital lab workstation included Adobe 
Photoshop CS2 for creation of derivative files, Adobe Acrobat for PDF 
conversion, and Datacolor Spyder 3 for display color calibration. 

The Pilot Project 

The principals involved in the Pilot Project were Michael Bennett, the 
new part-time Digital Initiatives Librarian at C/W MARS; Dodie 
Gaudet, Consultant for Bibliographic and Technical Services at 
CMRLS; and Jan Resnick, Assistant Regional Administrator for 
WMRLS. Dodie and Jan worked directly with the libraries in their 
respective Regions; Michael, who had previously attended the 



426 The Evolution of a Digital Repository in Massachusetts 

Northeast Document Conservation Center's (NEDCC) School for 
Scanning, did all of the scanning and was available to answer the 
technical questions. 

First some ground rules were established for the project. It was 
decided that the subjects of Industry and Agriculture would allow for 
maximum participation yet give the repository a focus. Concentrating 
on pre- 1923 objects would avoid copyright issues. Each library would 
be limited to 20 scans. Materials were to be two-dimensional, no 
larger than 11" x 17", and in good enough condition to be handled 
without falling apart. Once library staff had created the metadata, they 
would be responsible for transporting their items to and from the C/W 
MARS headquarters in Worcester. 

Information sessions were held in each Region to promote 
Digital Treasures and explain the Pilot Project. Next, Michael, Dodie 
and Jan developed an online questionnaire which was sent to the 
directors of all Regional member libraries. Directors were asked if 
they were interested in participating in Digital Treasures, how much 
material they owned, what kinds of materials they owned, how much 
they ultimately planned to contribute to Digital Treasures and related 
questions. 

As the cataloger, it was Dodie's responsibility to learn Dublin 
Core and develop a two-hour workshop. In preparation for the project, 
she attended Introduction to Digitization offered by NEDCC in 
October of 2005 and their Persistence of Memory Conference in 
November of the same year. Using the Dublin Core Metadata 
Initiative (DCMI) site (Dublin Core Metadata Initiative, 2008) and the 
Bibliographic Center for Research (then Collaborative Digitization 
Program) Best Practices Document (CDP Metadata Working Group, 
2006) she synthesized material on Dublin Core for Digital Treasures 
participants. 

Simple Dublin Core was used so as not to overwhelm the staff 
creating the metadata since most of them were not catalogers. The 
handouts are available on the Digital Treasures site. Workshops were 
scheduled in both the Western and Central Regions. Participants 



The Evolution of a Digital Repository in Massachusetts 427 

brought examples of the objects they planned to have scanned and 
practiced their Dublin Core descriptions on paper forms. 

The first library to complete the metadata for 20 images was the 
Jacob Edwards Library in Southbridge. The descriptions were keyed 
into an online form and e-mailed to Michael at C/W MARS. The 
photographs were delivered to C/W MARS for scanning and matched 
to the metadata. By the end of the Pilot Project in August 2006, 24 
libraries, twelve from each Region, each had a digital collection of 20 
images on Digital Treasures. Libraries also received a CD of their 
images to use for publicity or any other purpose they chose. 

To allow the repository reach beyond the Commonwealth's 
borders, Digital Treasures metadata have always been compatible with 
the OAI standard (Open Archives Initiative, n.d.). The metadata are 
harvested both by the Digital Commonwealth and OCLC's OAIster 
project. 

When specifications were being developed for Dublin Core 
records, the principals decided to use Library of Congress Subject 
Headings so that the records would integrate well with the MARC 
records in the online catalog. After the Pilot Project launched, Michael 
began working with staff at OCLC, C/W MARS and member libraries 
to crosswalk the Digital Treasures Dublin Core metadata into MARC 
format that could be loaded into OCLC Worldcat. As a subscriber to 
OCLC, C/W MARS would not only have their digital records available 
online, but would be able to export the records into their own public 
catalog with referring links back to the collections. Collections 
available in October 2007 were harvested and their MARC records are 
now available through WorldCat and the C/W MARS online public 
catalog as well as Digital Treasures. 

After the Pilot Project 

Up to this point, all costs for Digital Treasures had been assumed by 
C/W MARS, but the organization could not justify supporting this 
service exclusively with membership fees. The decision was made for 
C/W MARS to subsidize Digital Treasures, and charge participants a 
portion of the costs involved in scanning and maintaining the digital 



428 The Evolution of a Digital Repository in Massachusetts 

repository. With a vote of the membership, the following fee schedule 
was instituted and is still in effect: 

All C/W MARS members are allowed 20 free scans. Each scan 
beyond the first 20 is $8.50 (50% of the actual cost to digitize an 
object). After scanning, the library pays $i.oo/image annual 
maintenance fee. Libraries who are not C/W MARS members receive 
no free scans, pay $i2.75/scan (or 75% of the actual cost) and 
$i.50/image annual maintenance fee. 

The annual fee is assessed for the maintenance of each digital 
image's permanence over time through future platform migrations 
and regularly scheduled data refreshment cycles. It also covers 
maintenance of online delivery through CONTENTdm and the 
allowance of OAI metadata harvesting. 

Once the Pilot Project ended, activity in Digital Treasures slowed 
down drastically. A couple of additional information sessions were 
held along with another Dublin Core workshop. Everyone liked the 
idea of having photographs and documents from their local history 
collections digitized, but had difficulty finding the time to select 
objects and describe them in a useful way. Even though the financial 
situation was relatively healthy at the time, some libraries found they 
could not justify the scanning fees. In the fall of 2007, Dodie applied 
for and received a grant from the Greater Worcester Community 
Foundation (GWCF) to cover scanning fees for libraries. The grant 
was called the Central Mass. Memory Project. 

More importantly, the grant from GWCF funded a metadata 
specialist to work with the libraries selecting and describing their 
objects for digitization. Jeff Monseau, a Library and Information 
Science student at Simmons College who also worked part-time with 
the Archives of Mount Holyoke College, was hired as a "circuit rider." 
Because of the mission of GWCF, funds could only be used to help 
libraries within the Central Region. For those libraries, Jeffs 
experience, even the mere fact that he was available to help, was a 
tremendous boon and several new libraries added 20 images while 
other libraries increased the size of their digital collections. The grant 



The Evolution of a Digital Repository in Massachusetts 429 



funds enticed some non-C/W MARS members to participate in Digital 
Treasures which broadened the collection. 

In early 2008, Michael moved on to the University of Connecticut 
and Kristi Chadwick became the administrator of Digital Treasures. 

Digital Treasures on the Road 

Digital Treasures had been up and running successfully for over two 
years. Digital Commonwealth had developed a repository as well as a 
portal and was publicizing its services, stimulating more interest in 
digitization. For libraries that had not yet taken the step, the Technical 
Services Section of the Massachusetts Library Association (MLA/TSS) 
organized a one-day conference in the fall of 2008. The committee 
chose the topic Introduction to Library Digitization and invited 
several people from throughout New England to talk about their 
projects, how the projects were started and to describe the practical, 
day-to-day aspects. Kristi, Dodie and Jeff were asked to talk about 
Digital Treasures, its technical details, and the Central Mass. Memory 
Project. 

A Digital Barn Raising 

As one travels westward through the Commonwealth, the area 
becomes more sparsely populated. Most towns have libraries, but the 
libraries are often small and some are staffed by a single person. The 
median population in western Massachusetts is 1900. Springfield, the 
largest community, has over 150,000 residents, meaning some of the 
other communities are significantly below the median. The towns are 
also rather dispersed by New England standards, and there is little or 
no public transportation. The western Massachusetts approach to 
Digital Treasures - as with many projects - relies on team work. 

In May 2008, Digital Commonwealth received a grant from the 
H. W. Wilson Foundation for conversion efforts. The timing was 
fortuitous as WMRLS was in the process of organizing a "digital barn 
raising", a cooperative effort to get libraries started. To follow the 
theme, everyone contributed. Kristi Chadwick brought an 
introduction to Digital Treasures. A member of the Digital 



430 The Evolution of a Digital Repository in Massachusetts 

Commonwealth Board, contributed information about the Digital 
Commonwealth and described the grants available from the Wilson 
Foundation. A librarian from the Berkshire Athenaeum in Pittsfield 
reviewed the major copyright issues involved in digital projects. Dodie 
Gaudet presented the basics of Dublin Core and how libraries would 
use it. Member libraries brought a few of their local documents. Jan 
Resnick coordinated the program and arranged a few extra catalogers 
to assist in the process. 

The program consisted of presentations, questions and 
discussion, and concluded with a work session in which participants 
from member libraries began to develop Dublin Core metadata. This 
structure worked well for the group. They had the opportunity to ask 
questions (there were lots about copyright issues) and then had the 
chance to develop initial Dublin Core records for their documents in a 
supportive environment. WMRLS hopes to offer additional "barn 
raisings" expanding the skill base to a wider group of small libraries. 
Their collections contain a broad variety of treasures to be digitized. 
Buckland Historical Society, which participated in the barn raising 
session, went on to apply for and receive one of the Digital 
Commonwealth's Wilson Grants. Collaborating with the Buckland 
Public Library, they submitted over 50 objects to Digital Treasures. 
The Erving Public Library also received a Wilson grant and 
contributed 20 objects to the repository. The idea of the barn raising 
generated so much interest that participants were asked to contribute 
to a panel discussion at the day-long Digital Commonwealth 
conference held in the spring of 2009. 

Workflow 

Libraries with an interest in having objects digitized contact Dodie, 
Jan or Kristi. Items are limited to two-dimensional, 11" x 17" 
maximum size for scanning, or audio files in MP3 format. The original 
scope of agriculture or industry was lifted, so the theme of the 
collection is up to the library. For libraries that are not familiar with 
Dublin Core, Dodie assists with training, if based in the Central 



The Evolution of a Digital Repository in Massachusetts 431 

Region. Jan and Kristi assist libraries in the Western Region. This can 
be done by e-mail, phone, or in-person consultation. 

When the library has selected the objects they wish to add, Kristi 
gives them access to the password-protected online form. This form 
can be printed as a worksheet for initial descriptions, then entered 
online and submitted to Kristi by e-mail. The library also receives a 
copy for its own records. Once the library submits the first record for 
review, then all metadata can be sent. 

Once the metadata are received, an appointment is made to go to 
the C/W MARS headquarters in Worcester, where the scanning lab is 
located. Depending on the number of objects to be scanned, a two- 
hour or more block of time is reserved for the library. If the collection 
is new, preparation for the library's appointment will include setting 
up folders on the CONTENTdm server and workstation and verifying 
the administrative metadata, which are kept in a Microsoft Access 
database. 

When the library representative arrives with the images and/or 
documents, all items are matched by the identifier assigned by the 
library to the one listed in the metadata. Scanning on the Epson 
printer is done with its native program. All images are scanned in 
TIFF format for the "master" files, with a long side of 3000 pixels. 
Resolution of TIFF files range between 300 to 600 dpi, to achieve 
desired pixel dimensions. During the scanning, information from the 
metadata workform is copied into the MS Access Administrative 
Metadata database, including the title given to the object and the 
identifier assigned by the library. 

After initial scanning is complete, the TIFF files are manipulated 
in Photoshop and converted into the derivatives. This includes 
autoleveling and creation of JPEG files. Two separate JPEGs are 
created, a full-size image and one at access size for CONTENTdm with 
a long side of 600 pixels and 72 dpi. The full size JPEG is loaded to the 
CONTENTdm server directly, and the access images are loaded 
through the CONTENTdm workstation, where the rest of the 
metadata are loaded in from the workforms then uploaded to the 
server for web presentation. 



432 The Evolution of a Digital Repository in Massachusetts 

Stumbling Blocks 

From the beginning, there has been some resistance from a few 
libraries to the digitization of their collections. The reasons vary 
widely. Some staff and trustees are concerned that by making 
documents available digitally, people will no longer feel the need to 
visit their libraries. Only experience will breach this barrier. 

There is fear among a few people that digitizing library holdings 
and making them available on the World Wide Web exposes a library's 
materials to being used fraudulently. There is no evidence of this 
happening, but those who feel this way are not easily convinced 
otherwise. 

The copyright issue of post-1923 material is one that concerns 
library staff. Many photographs have been donated to libraries, and 
their origins are unknown. As Digital Treasures expands and libraries 
have more of their hidden collections scanned, this is an area that 
needs more research, more specific direction on what can and cannot 
be digitized and how to proceed if there is a complaint. In the 
meantime, concern over possibly breaking the law limits what objects 
libraries are willing to contribute to Digital Treasures. 

The digitization process has developed into a "three-legged" stool, 
comprised of Selection of Materials, Descriptive 
Cataloging/ Development of Metadata, and Financial Resources. 
Without support for all three, digitization does not happen. Carving 
out time for staff to select objects for scanning has become more 
difficult since Digital Treasures started. The world-wide financial 
crisis quickly trickled down to libraries and many libraries have had to 
lay off staff and reduce the working hours of others. Selecting 
materials for inclusion in Digital Treasures is time-consuming even 
when the local history collection is well organized and many are not. 

There are fewer people who feel comfortable with original 
cataloging. Even after objects have been chosen, it is easy to get 
bogged down in the process of descriptive cataloging. Digital 
Treasures participants have always been encouraged to create 
thorough metadata so that their images are more easily found via 



The Evolution of a Digital Repository in Massachusetts 433 

keyword searches. Unfortunately, the process is overwhelming to 
many, especially those who already feel time pressures. 

No one thinks that the scanning fees, which are subsidized by 
C/W MARS, are unreasonable. However, with tight budgets, every 
penny is scrutinized. For a C/W MARS member library, adding 20 
more images to its collection not only costs $170, but the annual 
maintenance fee is increased by $20 - a commitment that some 
libraries are unwilling to make. Financial incentives in the form of 
grants can help, especially when the grants include an experienced 
person to help with selection and metadata. 

The Future of Digital Treasures 

Currently the repository stands at 36 collections, with over 1300 
items. The digital repository's contents are searchable through the 
native Digital Treasures front end, the Digital Commonwealth portal, 
OAIster, major Internet search engines such as Google and Bing, 
WorldCat and the local C/W MARS online public catalogs. New digital 
file formats, such as audio recordings, have been included in the 
repository. The Digital Commonwealth presented a grant round in 
2009 to small institutions in Massachusetts, allowing some libraries 
to match funds to create new collections for the coming year. Another 
grant round is in the works for 2010. 

Knowing how much personal assistance with the selection and 
metadata process can benefit libraries, C/W MARS has added the 
Digital Treasures project to the University of Rhode Island (URI) 
Graduate School of Library and Information Studies (GSLIS) Site 
Experience database for a possible internship. While statistics show 
that Digital Treasures is accessed, with nearly five million hits and 
over 550,000 unique visitors in 2009, the success for the future lies in 
continuing contributions and promotion of the repository. 

References 

C/W MARS Inc. (2009). Connecting libraries in Central and Western 
Massachusetts. Retrieved on May 2, 2010 from 
http://public.cwmars.org/ 



434 The Evolution of a Digital Repository in Massachusetts 

CDP Metadata Working Group. (2006). Dublin Core metadata best 
practices, version 2.2.1. Retrieved on March 31, 2010 from 
http://vvrww.bcr.org/dps/cdp/best/dublin-core-bp.pdf 

Central Massachusetts Regional Library System, (n.d.). Homepage. 
Retrieved on May 2, 2010 from http://www.cmrls.org/. 

Digital treasures: A Central & Western Massachusetts digital library 
project. (2009). Retrieved on May 2, 2010 from 
http : / / dlib . cwmar s . org/ . 

Dublin Core Metadata Initiative. (2008). DCMI metadata terms. 
Retrieved on March 31, 2010 from 
http : / / dublincore. org/ documents / dcmi -terms / 

Open Archives Initiative (n.d.). Standards for web content 
interoperability. Retrieved on March 31, 2010 from 
http : / / www. op enar chives . org/ . 

Western Massachusetts Regional Library System. (2010). Western 
Massachusetts Regional Library System - Libraries Cooperating 
Communicating Sharing. Retrieved on May 10, 2010 from 
http://www.wmrls.org/ 



Collaborative Digitization Goes Local 



Ken Middleton, Mayo Taylor 
(Middle Tennessee State University) 



Abstract 

This chapter describes our efforts to build a digitization program in 
spite of limited funding. By obtaining small grants and forming 
partnerships with other campus units and institutions in the Middle 
Tennessee region, we have developed modest yet engaging collections 
and planted the seeds for additional partnerships. The article will also 
highlight the use of graduate students, efforts to streamline the 
digitization process, and lessons learned. 

Keywords: Collaboration, CONTENTdm, Collaborative digitization; 
Oral histories; Work flow, Community projects. 



Introduction 

In an era when many libraries are struggling to stay relevant to today's 
Facebook generation, digital projects have injected new purpose into 
Middle Tennessee State University's Walker Library. Staff members 
are excited about digitizing oral histories of pioneering women on 
campus, documenting the agricultural history of a neighboring 
county, or preserving seminal research on local rare plants. The digital 
collections under development by the Library have required 
collaboration with other campus entities and external institutions. 
Few of the digitized objects have come from the Library's own 
collections. In this chapter, we will describe the benefits of these 
partnerships and how they can be nurtured. We will also explore 



435 



436 Collaborative Digitization Goes Local 

many of the practical issues that new digitization programs face, such 
as securing staff, streamlining the digitization process, handling 
diverse file formats, and customizing collections. 

Efforts to collaborate with area institutions began as a natural 
extension of Walker Library's involvement in Volunteer Voices 
(http://volunteervoices.org), Tennessee's statewide collaborative 
digital project. MTSU librarian Ken Middleton served as the co- 
principal investigator of a three-year "Building Digital Resources" 
grant from the Institute of Museum and Library Services (IMLS) that 
coordinated the gathering of historical materials from organizations 
and institutions statewide. That project, based at University of 
Tennessee-Knoxville, focused on Tennessee history. Staffs were placed 
in each area of the state (west, middle and east) with a charge to 
collect unique photos, documents and other materials from cultural 
institutions. Selected items featured both major trends and figures in 
Tennessee history (e.g. early settlement, Ida B. Wells), and significant 
local items (e.g. Bradley County Courthouse). An important 
component of the project was to relate the digitized objects to the 
curriculum followed by schools in Tennessee. Walker Library was a 
satellite digitization center for the middle Tennessee region (Conner, 
Middleton, Carter, & Feltner-Reichert, 2009), with several librarians 
and staff participating. 

As the IMLS grant project ended, Walker Library began exploring 
ways that it could build its own digital skills while continuing to assist 
area archives, libraries, and museums to digitize unique collections. 
We would quickly lose skills and momentum if we did not build on our 
Volunteer Voices experience by securing the institutional support to 
initiate our own projects. Noting that a key mission of the University 
is to develop "mutually beneficial partnerships" (Middle Tennessee 
State University, 2002) we examined appropriate models among more 
experienced institutions. 

In 2002, The University of Pittsburgh's Digital Research Library 
began a two-year project to digitize a collection of 7,000 historic 
Pittsburgh images in partnership with the Historical Society of 
Western Pennsylvania, the Carnegie Museum of Art, and the 



Collaborative Digitization Goes Local 437 



University's Archives Service Center (Galloway, 2005). As evidence of 
the continued success of this approach, they have added three more 
partner institutions over the last five years and now have more than 
14,000 images in the collection (University of Pittsburgh Digital 
Research Library, 2010). The Digital Research Library has also 
collaborated with University of Pittsburgh faculty to create such 
projects as American Left Ephemera and Visuals for Foreign 
Language Instruction (Galloway, 2009). This program offers home 
pages for each of its 70 digital collections; each collection home page 
offers search and browse options, and a link to the partner 
institution's home page. 

IUPUI Library's Program of Digital Scholarship has collaborated 
with an impressive array of campus and local institutions. The 
collection websites and promotional material give every indication 
that the Program of Digital Scholarship treats these institutions as 
equal partners: the "Partners" page provides background information 
about each institution; logos of relevant partners appear on each 
collection home page; and a link for contact information appears in 
each record. Its list of partners continues to grow. Building on its 
initial collaboration with Conner Prairie Museum that features textiles 
from the museum's collection, the Library recently announced two 
Conner Prairie collections based on 3-D imaging technology. 

After touting the benefits of such collaboration (e.g., cost- 
effectiveness), Bishoff (2004) offers further guidance for librarians 
who want to forge lasting partnerships with other cultural heritage 
institutions: involve all partners from the start; respect each 
institution's mission and culture; and ensure that each partner 
benefits from the end product. 

Building on these models we outlined a vision for our own digital 
project design and received the support of our Dean to purchase a 
server, large flatbed scanner and a license to CONTENTdm software. 
We then leveraged that support to secure a $5000 grant from the 
University Special Projects Committee to fund graduate students for 
our first project, MTSU Memory. More on that collection will follow. 



438 Collaborative Digitization Goes Local 

Institutional Context 

Walker Library serves 25,000 students on a rapidly-growing campus. 
Its funding and staff levels are significantly lower than most 
comparable institutions within Tennessee and nationally - 
approximately 60% of the average of our designated peer institutions 
on a per-student basis in 2005-2006 (W. Black, unpublished data). 
Digital work has been accomplished by reassigning small portions of 
time for several librarians and one staff member, making extensive 
use of graduate students and purchasing equipment and software on 
an as-needed basis, usually at the end of the fiscal year when other 
needs have been satisfied. Digital projects management has been 
added to the load of our small staff. We have no programming support 
for customizing CONTENTdm or working with open source programs. 
This has hindered our ability to implement advanced features such as 
embedded audio and video players. 

Staffing 

The situation for digital projects did improve recently when the 
Library administration designated our working group as "the Digital 
Initiatives Team," thus formalizing our role in the Library, and clearly 
recognizing the roles of the team members: Ken Middleton, 
Reference/Microtext Librarian (Team leader); Mayo Taylor, 
Reference/Access Services Librarian; Fagdeba Bakoyema, 
Reference/Web Librarian; Lynda Duke, Microtext manager; and Jo 
Williams, Cataloging Librarian. 

Because these team members can only devote limited, irregular 
chunks of time to digital work and because they are drawn from 
different departments of the library, communication is the key to 
moving forward. We work hard to avoid problems by holding weekly 
meetings that also include our student workers, maintaining a wiki for 
the management of meeting minutes and other information, and by 
sending monthly reports of major activities to the Library Dean and 
Coordinators of both Library departments. 

We are able to balance the restrictions on our professional staff 
time by liberal use of graduate students from the Public History 



Collaborative Digitization Goes Local 439 



program on campus. They bring to the work an awareness of the 
historical and cultural significance of the collections, a desire to gain 
skills that are highly relevant to their discipline and a facility with the 
technology involved. Because many of these students will enter 
careers in history museums or archives they also make ideal liaisons 
to local historical agencies. Our graduate students have been 
supported primarily by small grants received in support of specific 
collections, as discussed below. We have supplemented this labor 
force with an undergraduate scholarship student from the Electronic 
Media Department, at no cost to the Library, and by utilizing students 
employed at the Microtext service desk for projects that can be done 
when they are not busy with patrons. 

Software /Hardware 

Staffing issues also heavily influenced our decision to purchase the 
CONTENTdm digital content management software. Although its 
shortcomings will be duly noted, CONTENTdm offers institutions the 
ability to quickly create digital collections with minimal systems 
support. In addition at least six institutions in Tennessee are using the 
system, creating opportunities for collaboration. 

The Library has purchased a good mix of digitization equipment. 
An Epson Expression 10000 flatbed scanner is large enough to handle 
most photographs and documents that we have selected for scanning. 
The Library recently purchased a Bookeye 3 planetary scanner with 
OPUS software. Intended primarily for the Interlibrary Loan 
department, it is available for digital projects outside standard 
business hours for fragile books and large format materials. We also 
have two portable scanners and a laptop computer to assist area 
institutions in digitizing their materials. 

Collections 

Campus Partnership: MTSU Memory 

In anticipation of MTSU's centennial year in 2011, MTSU 
Memory (http://library.mtsu.edu/digitalprojects/mtsumemory.php) 



440 Collaborative Digitization Goes Local 

will include photographs, documents, books, and oral histories that 
document our institution's growth from a small Normal School to a 
doctoral university with Tennessee's largest undergraduate 
enrollment. The Library's Special Collections hold many items of 
interest, including the institutional copies of yearbooks, faculty 
directories and the student newspaper. However, a larger repository of 
essential items is at the Albert Gore Research Center, which serves as 
a de facto campus archive. An early emphasis was to consult with the 
Director of the Gore Center to establish a working relationship and 
procedures for pulling and scanning materials. 

Collection Focus 

The digital projects team wanted to maximize usage of the 
collection by focusing on topical areas: the Great Depression, World 
War II and aftermath (e.g., GI Bill), the Vietnam War and student 
activism, and African Americans and women on campus. We expected 
the majority of items in the collection to be photographs, but wanted 
to include other types of primary sources (artifacts, newspaper 
articles, and oral histories) to present a multifaceted portrait of the 
University's history. The selection process for this collection has in 
fact been rather unusual because we initiated it without knowing what 
we would find, adjusting our expectations as new materials have been 
revealed. 

Early Implementation 

Walker Library launched its installation of CONTENTdm in late 
summer of 2007. Installation of the dedicated server and 
implementation of the software was handled by our Systems 
Librarian. Initial configuration of the CONTENTdm software was 
done by Ken Middleton when our first collection, MTSU Memory, was 
established. The remainder of the Digital Initiatives team members 
had been recruited and we had received guidance from our Dean on 
the direction that he would like us to take. Everyone on the team had 
some experience with digitization projects but none had worked with 
CONTENTdm. In an ideal world we would have enjoyed a beta period 
of testing the system and training the graduate students. In the real 
world, the clock was already ticking on a grant project and the 



Collaborative Digitization Goes Local 441 



graduate students needed work to do. They worked on the project 
more hours in most weeks than did their supervisors, leaving the 
professional team members hard pressed to hone their own skills to 
keep up with the graduate students. The students began selecting 
materials and scanning materials before any of us were proficient in 
using CONTENTdm and before we completed the configuration of the 
MTSU Memory field structure. One example of the consequences: as 
each record is created it is put in an approval queue to be reviewed 
and approved by one of the librarians. If the librarian isn't able to keep 
pace with the student, simple errors are repeated many times before 
they are corrected, which exacts a toll in terms of time wasted on 
corrections. 

To slow down the rate at which items were loaded into 
CONTENTdm we gave the students guidance on the types of items we 
were looking for and sent them over to the Albert Gore Research 
Center. They concentrated on selecting items, but without the benefit 
of prior experience in completing metatadata they didn't know exactly 
what information they needed to capture to assure complete 
documentation of the items. The result was that many objects were 
loaded with incomplete data and improper form. Predictably, we are 
still returning to some of the earlier records to fill gaps and fix 
inconsistencies. 

By summer of 2009 we again posted a job notice with several 
graduate programs on campus and again selected two Masters 
candidates from the Public History program. This time we started 
each of them with an introduction to the collections they would work 
with and the metadata guidelines available on our wiki followed by 
CONTENTDM tutorials and documentation. Finally, they began by 
importing and creating metadata for objects selected by our previous 
workers. Only after these initial projects were complete were the 
students free to find new materials for the collections. 

Expanding Opportunities 

Our original vision for MTSU Memory was a collection composed 
primarily of photographs and documents. However, we adjusted that 



44 2 Collaborative Digitization Goes Local 

concept when given the opportunity to include excerpts of audio and 
video interviews and oral histories. 

Audio materials: As our graduate students investigated 
materials at the Gore Research Center they discovered an excellent set 
of oral histories that had been conducted with alumni from past 
decades. From that extensive collection, twenty-five were selected 
because of their relationship to the lead themes of our collection, such 
as World War II, Civil Rights, etc. Our initial plan was to digitize and 
display the typed transcript of each interview and then offer selected 
audio excerpts as a supplement. When a new Director joined the Gore 
Center he decided that the transcripts needed additional work to be 
ready for online publication. We adapted by finding a work around to 
offer only the selected audio excerpts. It consisted of loading MP3 files 
to our server, but not within CONTENTdm, importing a photo of the 
subject as the digital object for the CONTENTdm record, and then 
creating a metadata field to include links to the MP3 files. The work of 
selecting, reviewing and excerpting the oral histories was done by one 
of our graduate student workers as an internship for course credit. She 
selected and downloaded the freeware program, Audacity, to manage 
the sound files and participated in designing the customized process 
for creating the records. We hope to include the entire transcript of 
each record at some point. At present our CONTENTdm records direct 
users to the items available at our partner across campus. 

Video materials: Another treasure trove at the Gore Center was 
a collection of interviews that were broadcast on the campus television 
station in the 1980s and 1990s. Of particular interest was a series of 
four lengthy interviews with Miss Mary Hall, an Education professor 
important to MTSU and the related University demonstration school 
from 1929 to 1963. We again turned to one of our graduate students 
who had recently completed a course in documentary filmmaking. She 
was able to use her personal computer (a MacBook) and software 
(iMovie) to create a series of video excerpts that encapsulate 
important phases of Miss Hall's life. 

Book-length publications: As we researched our own campus 
for the MTSU Memory project, we frequently consulted The First Fifty 



Collaborative Digitization Goes Local 443 



Years, a campus history written by alumnus Homer Pittard, that 
covered the period 1911-1961. Because MTSU held the copyright for 
the book we decided to digitize the entire volume. This initiated an 
intensive learning experience about the pros and cons of compound 
objects and ultimately the process of creating complex PDF files. The 
process of scanning the pages and doing OCR through the Abbyy Fine 
software was quite straightforward. A definite learning curve was 
encountered in figuring out how to structure a compound object from 
the hundreds of page images. However, once the book was loaded into 
the MTSU Memory collection, we were dissatisfied with the results, 
largely because queries could return dozens of individual pages 
intermingled with other materials. In addition, when we figured out 
the compound object expense in terms of licensed objects used, we 
rethought our original approach. In brief, a 200 page book exacts a 
cost of 201 objects utilized, 200 for the images and one for the 
combined record. By comparison, a book imported as a PDF made 
from the same original image files costs only one licensed object. 
While this may seem an arcane concern, with a licensing level of only 
10,000 items we could envision our capacity diminishing quickly if we 
continued to do book-length items. So, The First Fifty Years was 
reloaded as a PDF as was our second book, The Raider Forties, and a 
thesis. Smaller multi-page items continue to be treated as compound 
objects. 

It is worth mentioning that the experience of the user is quite 
different for audio, video and large PDF text files in our collections 
when compared with objects loaded as single JPEG files, compound 
objects or small PDFs. Whereas those files provide a display image 
immediately upon opening the record, the other file formats all 
depend upon software programs external to CONTENTdm (e.g. Adobe 
Acrobat or Quicktime) to open the file. In the case of large media files 
or PDFs the process can be relatively slow. The record display is also 
marked by the rather unimaginative default message of "Access this 
Item," which is not far removed from the archaic "click here" message 
of early web pages. We hope to find a partial solution to this problem 
through embedded media players. 



444 Collaborative Digitization Goes Local 

Customizing Our Flagship Coiiection in CONTENTdm 

Although CONTENTdm is an excellent tool for creating digital 
collections quickly, it does not provide a particularly user-friendly 
public interface. All searches retrieve results from all collections, 
which can be problematic when you have collections as varied as 
university history and ecology. We have created a separate entry page 
for each collection with a brief description of the collection, a search 
box, a link to an advanced search, and browsing options. 

The lack of default browse options for subjects, time periods and 
formats has been particularly frustrating. CONTENTdm requires us to 
create the browse lists manually by running "custom queries" for 
every subject, for instance, that appears in a list. This process requires 
a significant amount of time. We have recently discovered Elias Tzoc's 
web-based tools for creating CONTENTdm browse lists (Tzoc, n.d.). 
Creating the list of 200 subject headings in MTSU Memory would 
have taken approximately 10 hours if we had used custom queries. 
After learning a few fairly simple steps, we created the list in just a few 
minutes using Tzoc's tool. 

One extremely useful feature of a recent CONTENTdm upgrade is 
that search results now include facets that help users narrow their 
searches. For example, a simple search retrieves a list of records that 
can be narrowed by time period, subject, and genre. However, making 
changes to the default facets required too much time. Rather than 
simply check the desired facets in an administrative module, one has 
to find the relevant line of code in a PHP file, use correct PHP syntax, 
and determine the correct code for each facet. Many libraries may lack 
the technical staff to make such changes. 

We are making efforts to attract users through additional access 
methods. Our metadata librarian has set up a profile in OCLC's digital 
collection gateway to streamline the creation of WorldCat records for 
our digital items. We are working at the state level to use the 
CONTENTdm Multi-site Server, which would create a central site for 
searching across CONTENTdm collections in the state. 



Collaborative Digitization Goes Local 445 



Community Partnership 

During 2009 the Digital Initiatives team received a $1000 grant 
from the MTSU Public Service Committee to assist several community 
associations with digital projects. Selected organizations were invited 
to propose projects; the first application received was from the Arts 
Center of Cannon County (ACCC). They had been given a collection of 
nearly 1000 photographs that were taken between 1944 and 1977 by 
William L. Clements, District Conservationist for the Soil 
Conservation Service (SCS). He served Cannon County, a small 
agricultural community just east of Murfreesboro and MTSU. The 
photos documented the farming methods that were disseminated by 
the SCS in that period. Significantly, each photo had descriptive 
information typed onto the image back, which made the task of 
describing each image quite easy. 

The Clement Collection was much larger than our initial idea of 
what would be handled for each project with our small grant, with 
nearly 1000 photos available, but it was so compelling that we worked 
to find a way to accept it. The answer came through a supplemental 
$300 grant from our local Gannett newspaper and through use of a 
student worker who needed to fulfill work hours for his scholarship. 
The Cannon County, Tennessee Collection (http: //library. 
mtsu.edu/digitalprojects/cannon.php) includes approximately 400 
photographs selected from the total available. 

We accepted the Clement photos because they beautifully 
illustrate and describe the details of the County's predominant 
occupation, farming, over a period of three decades. In addition, the 
collection was an ideal test case for working with area institutions to 
create digital collections: 

• The Arts Center is an essential institution in a community where 
farming is still economically and culturally significant.. Many of 
the families pictured are still on their farms today. 

• The Arts Center staff had already worked on a promotional plan 
for the collection and have developed plans for creating an exhibit 
based on the photos. 



446 Collaborative Digitization Goes Local 

• Additional rich contextual material is available. Three hours of 
recent oral history interviews with the photographer will be 
included along with an interpretive essay from Evan Hatch, 
folklorist with the Arts Center. 

• There were no copyright issues because the photos are the product 
of a federal employee. 

• The photos have been relatively easy to scan and process because 
the photographs are all the same size. Moreover, detailed 
metadata is already recorded on the back of each photograph. 

The homogeneity of the collection also provided an opportunity to 
develop streamlined processes, for example by loading nine images at 
a time onto the large format scanner, then utilizing Photoshop tools to 
deskew and adjust each image before saving it individually. We also 
were able to make optimum use of the template option within the 
CONTENTdm Project Client to automate the entry of identical 
information into selected metadata fields as each object is imported. 
In this manner large groups of records could be created in the Project 
Client and then uploaded in a batch operation for approval. The 
approval process itself was efficient because only the relatively few 
unique fields in each record (e.g. title, LC subject heading) needed to 
be carefully checked by the approver. 

We have also used Elias Tzoc's web-based tool to create a tag 
cloud based on subject headings in the collection. Tag clouds offer a 
useful alternative to traditional subject heading lists because they 
provide users with a quick sense of the major subjects in a collection. 

The Digital Initiatives team has made every effort to involve the 
Arts Center of Cannon County in developing the collection. Initial 
meetings with Evan Hatch, the Center's folklorist, covered everything 
from selection issues to scanning standards. The Center's logo appears 
on the collection web site, and links to the Center's web site. Treating 
the Center as an equal partner improved the quality of this initial 
digital collection, and it has also increased the likelihood of future 
partnerships. Early success of the project has encouraged discussions 
of additional materials that may be forthcoming to document the 
history and deep crafts tradition of Cannon County. 



Collaborative Digitization Goes Local 447 



New and Forthcoming Projects 

Buchanan Collection - will document the history of a local 
family that has included a Nobel Laureate MTSU alumnus, a 
governor, and an early settler of Nashville. The collection is being 
compiled in cooperation with family members. 

Cedar Glades Studies Collection - a partnership with the 
MTSU Center for Cedar Glade Studies will digitize research on the 
rare cedar glade ecosystems of Middle Tennessee. Most notably the 
1948 doctoral thesis of Dr. Elsie Quarterman was digitized with 
support from a small grant passed through the Biology Department. 
Her thesis was handled by two practicum students from the UT- 
Knoxville School of Information Science. 

Shades of Blue and Gray: Reflections of Life in Civil War 
Tennessee - will include digital images of 250 artifacts from small 
archives, libraries, and museums across the state. In partnership with 
Celia Walker, director of Vanderbilt University's Peabody Library, we 
will create a metadata template in CONTENTdm, enter metadata, and 
customize the CONTENTdm database for items collected by other 
entities. 

Stones River Battlefield Historic Landscapes - is digitizing 
deeds and other materials to document the largely forgotten African 
American community of Cemetery, which developed after the Civil 
War on land that would become the Stones River National Battlefield. 
Created by MTSU history professor Rebecca Conard and her graduate 
students, it is hosted by the Library and Digital Initiatives staff have 
played an advisory role. 

Lessons Learned 

The Digital Initiatives team's collaborative approach has made it 
possible to take on more projects. In turn, our list of lessons learned is 
comparatively long and keeps growing as we face challenges specific to 
new projects. However, as the list below illustrates, many of the 
lessons apply either to every digital project. 



448 Collaborative Digitization Goes Local 

• Libraries that lack a digital projects department (and the 
corresponding budget line) should consider applying for small 
grants. In our case, the three grants have been enough to fund 
students doing the scanning and metadata entry. 

• Recruit student workers / interns with relevant skills, and offer 
them a rewarding work experience. Recognizing that our public 
history interns are indispensable to the success of our digital 
projects, we have treated them as equal partners on the Digital 
Initiatives team. We have encouraged them to attend our weekly 
meetings, welcomed their suggestions for the improvement of 
digital projects, and tailored their internships to their interests 
and skills as much as possible. 

• Be flexible. One of our best decisions has been to include 
interviews in audio and video formats in MTSU Memory, in spite 
of production delays. In addition, we adjusted our very modest 
approach to working with area institutions when we saw the 
potential of working with the Arts Center of Cannon County. 

• An effective communication structure is essential, particularly for 
new digitization programs. By having weekly meetings, we have 
addressed metadata and technical problems before they become so 
large that they are difficult to correct. 

• Communication is important when working with collaborative 
partners; tact, patience, and ability to explain both the process and 
benefits of digitization can not only help put partners at ease, but 
also pave the way for additional joint projects. 

• Invest time in strategic planning. By focusing so much on the 
details of specific projects, we have not devoted sufficient time to 
plan for a robust approach for digital preservation. 

Conclusion 

Perhaps the most important lesson learned mirrors our experience 
with the statewide Volunteer Voices project. Small libraries, archives 
and museums have valuable collections, and often want assistance in 
digitizing these collections. By sharing our successes and growth pains 



Collaborative Digitization Goes Local 449 



in working with an academic archive, a small arts center/museum, 
and a growing list of other partners, we hope to encourage other 
institutions to proactively seek out collaborative digitization 
partnerships at the local level. 

References 

Bishoff. L. (2004). The collaboration imperative. Library Journal 129 
(1), 34-35- 

Conner, T.R., Middleton, K., Feltner-Reichert, M., and Carter, A. 
(2009). Volunteer voices: Tennessee's collaborative digitization 
program. Collaborative Librarianship, 1 (4), 122-132. Retrieved 
from http://collaborativelibrarianship.org/index.php/jocl/ 
article/view/ 38/26 

Galloway, E. (2005). Historic Pittsburgh image collections. D-Lib 
Magazine, 11 (11). Retrieved from: http://www.dlib.org/dlib/ 
november05/nfeatured-collection.html 

Galloway, E. (2009). Challenges of creating digital libraries: 
Digitizing, organizing, storing, accessing content. 2009 CSE 
Annual Meeting, May 5, 2009. Retrieved from: 
http : / / www.councilscienceeditors. org/ events / annualmeeting09/ p 
resentations/galloway.pdf 

IUPUI University Library. (2009). Library grant makes more 
Indiana history accessible to all. Retrieved from http://www- 
lib.iupui.edu/ node/ 1306 

Middle Tennessee State University. (2002). Academic master plan: 
Middle Tennessee State University 2001-2012. Murfreesboro, TN: 
Author. 

Tzoc, E. (n.d.). Re-Indexing CONTENTdm metadata. Retrieved from 
http://staff.lib.muohio.edu/~tzocea/files/CONTENTdm/re- 
indexing/ 

University of Pittsburgh Digital Research Library. (2010). Historic 
Pittsburgh image collections. Retrieved from: 
http://images.library.pitt.edu/pghphotos 



Picturing the Past and Planning for the 
Future: Central Florida Memory 



Lee Dotson, Selma Jaskowski 
(University of Central Florida Libraries) 



Abstract 

Central Florida Memory is a collaborative digitization project to create 
a collection of primary sources related to Central Florida's history that 
provides a compelling look at Central Florida before the area became a 
tourist destination. Through cycles ranging from limited resources to 
ample grant funding, the collaboration has grown from a small pilot 
project with three partners to over 80,000 images contributed by 
seven partners. Organized for the user, the website has been enhanced 
to create more user friendly methods of access and includes storylines, 
lesson plans, and links to other sites of interest. This paper provides 
lessons learned about how to incorporate a variety of materials from 
diverse institutions into a single digital collection. 

Keywords: Academic library, Bethune-Cookman University, Central 
Florida Memory, Collaboration, CONTENTdm, Digitization, 
Historical Society, IMLS, LSTA, Museum, Museum of Seminole 
County History, Olin Library of Rollins College, Orange County 
Library System, Orange County Regional History Center Public 
library, Stetson University, University of Central Florida Libraries. 



The Beginning - 2002 

When the directors of the University of Central Florida Libraries and 
the Orange County Library System came up with the idea for Central 



450 



Central Florida Memory 451 



Florida Memory (CFM) over drinks in 2001, it seemed like a friendly 
way for the two institutions to collaborate on a project that would 
serve both academic and public library patrons and foster inter- 
institutional understanding. 

With the addition of the Orange County Regional History Center, 
the three institutions had the necessary resources to begin building an 
exciting digital collection. In 2002, the group began meeting to 
evaluate the collections of the three partners and determine what 
materials in each collection would provide the most compelling and 
engaging look at Central Florida's history. 

Coming from very different missions and working environments, 
the partners recognized early on that this collaboration would take 
some work. Staffing, funding, and the capacity for participation 
needed to be discussed and an understanding of each institution's 
identity and purpose gained. Through quarterly meetings where 
brainstorming was center stage, the group identified the strengths 
each partner could bring to the project. Committees were formed and 
populated with members from each institution to foster 
communication among participants as well as organize functional 
responsibilities. While the full group discussed and agreed on overall 
structure and support, the committees focused on operational tasks 
such as selection, metadata and cataloging, storyline, marketing, and 
evaluation and assessment. The scope of the project was defined, 
collections examined and material chosen for inclusion in what was 
hoped would become the definitive site on the history of Central 
Florida. 

Pilot Project - 2002 

With the goal of applying for grant funding, the partners conducted a 
pilot project to prove the viability of such a collection. It would also 
serve to lay the foundation for scanning materials at a centralized 
location and create a website dedicated to Central Florida's history. 
Since their vision included building a digital collection with 
contributions not just from the original three partners but from 
libraries, historical societies, and museums throughout Central 



45 2 Central Florida Memory 



Florida, the pilot project would be the beginning of a regional effort to 
expose hidden collections in small institutions. By digitizing materials 
from organizations with wonderful collections but no digitization 
resources, Central Florida Memory would be the vehicle to make these 
rich resources available electronically. 

To begin, materials created from 1880 to 1930 were selected from 
the collections of all three institutions. The UCF Libraries offered their 
digital imaging facilities and servers, Orange County Library System 
lent the services of their graphic designer, and the Orange County 
Regional History Center contributed an abundance of material to the 
collection. The partners agreed to deliver materials for scanning to the 
Digital Services unit at the University of Central Florida Libraries. 
With one part-time scanning assistant, a flatbed scanner, and a digital 
camera, work began on digitizing thirty-five items totaling 
approximately 2,000 images from materials such as maps, letters, city 
directories, monographs, and ephemera. Overall, Central Florida 
Memory adhered to the recommendations set forth by the Publication 
of Archival, Library, Museum Materials (PALMM) and Cornell 
University. PALMM guidelines were routinely referred to for 
copyright concerns and digitization standards (PALMM 
Documentation, 2010). In addition, Cornell's Moving Theory into 
Practice Digital Imaging Tutorial provided a strong foundation for 
making decisions regarding image creation, file management, quality 
control, and digital preservation (Moving Theory, 2010). Size, 
condition, content, and intended end use of the item to be digitized 
were several of the factors taken into consideration when determining 
the most appropriate method and equipment. 

Since the materials provided by the Orange County Regional 
History Center and the Orange County Library System were often 
fragile or rare materials, an environment of trust was essential to the 
success of the project. This was especially important for the History 
Center as it was not ordinarily their practice to allow materials to 
leave their reading rooms much less travel across the city. 
Representatives from the partner institutions were invited to spend a 
day with the scanning assistant to alleviate any concerns about the 
care and handling of materials as well as have a front row seat to the 



Central Florida Memory 453 



scanning process. With the small number of items involved in the 
pilot, tracking materials was easily performed using spreadsheets and 
email, and delivery and return of items was handled in person. 

The next step was to choose a content management system for 
storing, searching, and retrieving the digital images. CONTENTdm 
was chosen because it could either be set up rather quickly and used 
easily by all levels of staff as an out-of-the-box installation or 
customized to create a unique look and feel for the collection (Lavoie, 
2005). During the pilot, the installation was customized and objects 
were loaded into three distinct institutional collections. When 
contributing materials, partners were asked to fill out forms 
describing their materials. This information was used to create the 
metadata when the images were added to the collection. With a 
CONTENTdm collection of close to 2,000 images and a website in 
production, Central Florida Memory was born. 

In order to present the collection in a way that was most 
appealing to a wide audience, "stories" were included to provide 
background for the images found there (Success story, 2009). The 
storylines were divided into sections - Dreams and Schemes; Roads, 
Rivers and Rails; and Critters, Crackers and Cottages. Each of these 
sections was divided into subsections and included topics like tourism, 
health, hotels, railroad, water, roads, nature, social life, people, and 
place. In these subsections users would learn why people moved to 
Central Florida, how they traveled, and what they experienced in this 
new, harsh climate (Gronlund, 2005). Users could search and browse 
the collection itself or get to images of interest through the stories. 

The logo for the project and original website design were 
contributed by the Orange County Library System. The pages existed 
separately from CONTENTdm itself and supported the Florida 
Stories, Teachers, Site Map, Contact Us, and Links parts of the site. 
The link to Collection and any links in the other sections to images in 
the collection took the user from the website to CONTENTdm. 



454 Central Florida Memory 




TRAL FlORID 



Follow our path and listen to the stcriet of Ci(/<ji, T 
Alligators, Space, and more. Travel to the 
maps, steamship tickets, postcards, 
diaries. Hear Helen Witney writing to her me 
In England about ehs joys and trials of life in Writer Park, and 
Narcoose-fij In 1385. Read about an "Alligator Hunter In the 
muting" in 1909- "Gimme • gun quick I u«n ■ 'g*tor!' a «i 
Buddy « he tumbled dawn the <gmp*nlcwti«iy." m 

'foil 

G«t tO know Central Florida today, txploro yesterday., ind 
crwlr a historic tomorrow. 




Figure CFM-1. Central Florida Memory Website, 2003 (Central Florida Memory, 
2003) 

On 21 May 2003, the Central Florida Memory website and 
collection were unveiled at a reception at the Orlando Public Library. 
After presentations describing the Central Florida Memory concept, 
how it was created, and plans for its future, refreshments were served 
and attendees were shown how to navigate the collection on multiple 
workstations set up in the meeting area. Included in the audience 
were representatives from area libraries and historical societies, as 
well as members of local government. Response to the new digital 
collection was very favorable. Contact information for potential future 
contributions was offered by several attendees. 

Upon completion of the pilot project, the group applied for a 
Library Services and Technology Act (LSTA) grant but was 
unsuccessful. A year later, the group applied for and received a 
National Leadership Grant from the Institute for Museum and Library 
Services (IMLS) to expand the collection. The National Leadership 



Central Florida Memory 455 



Grant was a perfect fit for Central Florida Memory given that the 
grant's goals included supporting "projects that have the potential to 
elevate museum and library practice" while advancing "the ability of 
museums and libraries to preserve culture, heritage and knowledge 
while enhancing learning" (IMLS National Leadership, 2010). As part 
of that grant project, additional partners were invited to join. The Olin 
Library at Rollins College in Winter Park, Florida and the Museum of 
Seminole County History in Sanford, Florida were welcomed to 
Central Florida Memory. 

As part of the IMLS grant proposal, a formal mission statement 
was developed to illustrate the partners' ongoing commitment to 
building Central Florida Memory. 

The mission of Central Florida Memory is to create, through collaboration among 
museums, libraries, historical societies, and archives in Central Florida, a 
comprehensive digital collection, serving the needs of scholars, students, teachers, 
and residents. Here they will access textual and graphical images and objects that 
document the region's historical and cultural heritage. 

In order to fulfill its mission, CFM established the following goals: 

Provide access to the rich and varied content of libraries, museums, historical 
societies, and archives throughout Central Florida. 

Create an extensive digital resource for scholars, students, teachers, 
genealogists, Central Florida residents, visitors, and the general public, 
promoting lifelong learning. 

Assist museums, historical societies, and libraries in digitizing their collections 
and become a forum for sharing expertise and technical knowledge. 

Offer a digital archive to those museums, historical societies, and archives 
unable to establish their own digital infrastructure. 

Provide an innovative digital resource for K-12 teachers and parents to bring 
alive the region's history into existing and new lesson plans. 

Expand, enhance, and disseminate the collaborative structure model 
developed during the CFM pilot phase. 

Contribute to the larger body of knowledge of Florida history via access to the 
CFM website through other statewide project sites. 

From the narrative of the IMLS grant submitted 3/31/2004 (Central 
Florida Memory Grant Writing Committee, 2004). 



456 Central Florida Memory 

IMLS Grant - 2004 

The IMLS grant phase provided a tremendous boost for Central 
Florida Memory. The funds received from IMLS made possible a 
commitment to increase the collection by up to 50,000 images. 
Additional activities included creating lesson plans to support K-12 
education, expanding and enhancing the storylines and website, 
marketing the collection through promotional materials and a video, 
and presentations and poster sessions at conferences. At this juncture, 
deciding how to organize the collection was critical in affecting how 
the end user would search and retrieve items. Should we continue to 
divide the materials based on the contributing institution, by some 
other criteria, or not at all? 

CONTENTdm allows creating multiple collections based on 
locally-determined parameters, or the ability to have one large 
collection and allow metadata to create subsets of materials for users 
as they browsed or searched. Having created three distinct collections 
as part of the pilot project, we knew that there could easily be 
confusion in searching and difficulties in upgrading. In the pilot 
model, each time we upgraded CONTENTdm, it had to be performed 
3 times, once for each collection. With the addition of 2 new partners, 
it would be 5 times. Given those issues and understanding that the 
original intention of Central Florida Memory was to be a collection of 
a variety of materials representing all institutions, we combined the 
institution-specific collections into one single collection. This would 
make searching across all formats and all collections simpler and 
applying upgrades less complicated. However, we soon found out that 
combining collections had its own challenges. Working with a wide 
variety of materials in a single combined collection would impact 
production processes and necessitate new approaches. 

With these changes and an ambitious workload ahead, some 
adjustments to the committee structure and work processes were 
needed. A Steering Committee was formed of the directors of the 
partner institutions for administrative purposes. An Operations 
committee was composed of institutional representatives who then 
served on one or more of the seven working groups. 



Central Florida Memory 457 



The Curators Working Group was responsible for selecting 
materials for digitization. Taking our initial cues from the pilot 
project, materials were tracked on a shared spreadsheet and itemized 
lists accompanied delivered materials. It was soon discovered that one 
of the differences between scanning 2,000 images and 50,000 images 
was how quickly Digital Services became inundated with materials 
lacking descriptions or a process for tracking their progress (Dotson, 
2005). Boxes of loose photographs seemed to appear from out of 
nowhere and books with similar titles were easily confused. A 
Digitization Request Form was created and partners were responsible 
for submitting a form for each item and arranging for delivery of the 
materials to the University of Central Florida Libraries Digital 
Services unit (Digitization spec kit, 2005). The form helped to 
consistently track materials through the use of uniform bibliographic 
information and a computer generated unique identification number. 
It allowed for entry of information pertaining to the lending 
institution's contact information; physical item information such as 
title, author, subject, description/notes, date original, publisher, 
number of pages, format, call number/identification number, and 
storyline; and digitization information including number of images to 
digitize and notes regarding special handling or other circumstances. 
The lending institution received a confirmation page upon submission 
of the request form to be printed and delivered along with the 
appropriate item. The confirmation page and unique identifier 
accompanied the item throughout the process from material inventory 
check in to scanning to metadata. In addition to providing a well 
organized tracking system, these measures resulted in a meaningful 
transition from a physical item to a digital image. 

It was soon discovered that the scanning, metadata, and upload 
processes for some of the materials varied from the normal 
production processes. As general practice, when materials were 
received they were reviewed to determine the appropriate equipment 
to be used for digitization. The item was scanned, metadata created, 
and the images and accompanying metadata uploaded to 
CONTENTdm. However, special requirements for some materials 
necessitated a "projects within a project" approach to be developed. In 



458 Central Florida Memory 



all, five different "projects" (as discussed below) with unique 
characteristics were completed and added to the collection. 



D Luteal 5*rtK**DjMbM* [\»t\f1i*i«tc*t} 



J» w 9" ft—* w e—fc i«* -a**- a* ****** 




:'*y.r«j.^ 



T„(,l Hou'5 



Apr 2DDt Apr 



US 

1 il-.J- .i^js r> 



Sun | Mm| In | VM| Ttm 


rn 






« 








I 




H 






J. 


S 






H 









Figure CFM-2. Screenshot of the Digital Services Tracking Database (Digital 
Services, 2005) 



Newspapers 

The challenges presented by including the newspapers in a 
collection of mixed materials included inconsistency in the physical 
size, large format scanning, and optical character recognition for full 
text search capabilities. While several newspapers were digitized 
during this time, the most extensive experience was with the 
Sandspur, the student newspaper of Rollins College. The Sandspur 
began as a literary journal in 1894. Digitization of the first seventeen 
volumes proceeded normally as each issue ran 20 to 40 pages and 
measured in size from 27x20 cm to 23x15 cm. With volume 18, the 
format changed to that of a more typical newspaper length and size. 
These fragile bound newspapers measured anywhere from 38x27 cm 
to 41x28 cm and required scanning on a Kaiser rePro graphic stand 



Central Florida Memory 459 



with Karden-Linhof camera. Additional time and attention to detail 
were required to complete the digitization of the newspapers. Once 
scanned, the images were loaded into CONTENTdm as jpegs. Lacking 
OCR capabilities at the time, the metadata cataloger painstakingly 
created page level metadata to describe the contents for each image 
(Dunlop, 2005). 

Funeral home records 

The Undertaker's Memoranda from the Carey Hand Funeral 
Home records was an important collection used extensively for 
genealogical and historical research on Central Florida's people and 
places. The collection contained records and ledgers on services 
provided at the Carey Hand Funeral Home from 1891-1955. Forty-four 
books of the Undertaker's Memoranda covering the time period from 
June 15, 1891 through November 10, 1927 were made available on 
Central Florida Memory. In order to make the records searchable 
online, the handwritten fields had to be transcribed and linked with 
the appropriate image. A set was created for page level, field-by-field 
metadata, to represent information about the person, death, family, 
and burial. Efforts were coordinated between Digital Services and 
Special Collections so that as the books were being scanned, 
information from the handwritten fields was entered into a 
spreadsheet. Prior to uploading the materials, the appropriate rows on 
the spreadsheet were turned into text files that contained the 
associated metadata for each page. The end result was a collection of 
records searchable by name, date of death, cause of death, age when 
deceased, place of residence, gender, race, parents' names, and burial 
place. 

Voter's registration records 

The Orange County voter's registration records collection was 
also of great interest to genealogists. It contained the voting records of 
registered voters in Orange County cities from 1912-1934. All ledgers 
were numbered as volumes. Each volume contained the voter's name, 
age, race, occupation, local residence, party affiliation, whether a poll 
tax had been paid, and nativity or place of birth. Other information 
may have included the voter's declaration of naturalization, date of 



460 Central Florida Memory 



voter's registration with a political party, and remarks concerning 
disabled veteran voters, registration changes and deceased voters. 




Figure CFM-3. Carey Hand Funeral Home Undertaker's Memorandum ("Book 24: 
Carey Hand," 2005). 



The voter's registration records were slated to be digitized from 
microfilm to reduce the scanning workload. Unfortunately, the only 
microfilm copy available had received a great deal of use in a public 
library resulting in unreadable poor quality images. The original 
records had been in storage for many years and they had been 
randomly stored in boxes. It took several days to sort and organize the 
books by voting precinct before scanning or metadata creation could 



Central Florida Memory 461 



begin. A graduate student was hired to assist with metadata for the 
257 books that contained over 8,000 records. Since the records were 
in an oversized ledger book format that read left to right across both 
pages, it was quickly realized that scanning procedures would need to 
be adjusted. Using the overhead scanner proved to be too time 
consuming for the 8,000 plus images. 



rf<M,FLOfiiw Mi .mops 




Figure CFM-4: Magnified image of Orange County Voter's Registration Record 
("Orange County Voter Registration," 2005). 



The most efficient method of scanning these images was 
determined to be scanning the left and right pages individually and 
manually merging the two into one cohesive image. For each of the 
voter's registration images available, 3 images were created. The end 
result of zoomable jpeg2000s allowed the user to see an overall view 
of the page and zoom in to read the information line by line. 



462 Central Florida Memory 

Handwritten letters 

Two sets of handwritten letters, providing great insight into the 
lives of settlers in early Central Florida, became the fourth "project." 
The first set was penned by Arthur William Catesby Smyth. Smyth 
wrote letters to his fiance and first cousin, Beatrice Alice Smyth, while 
she was living in England and he was living in Orlando. Smyth was a 
partner in Smyth and Co. Orange Packers and Shippers. The letters 
focus on the developing relationship of Arthur and Beatrice, but also 
include some general information on the citrus industry, property 
prices and social events. The other set of letters was written by a 
newlywed transplant to Florida, Helen Warner, who regularly wrote 
home to her mother, Mrs. Haig, who lived in Buckinghamshire, 
England. The letters describe both personal matters, and those of 
larger interest, including the establishment of an orange grove and the 
development of the city of Narcoossee. Helen and Bill Warner moved 
to Florida around 1885 and stayed in Winter Park while waiting for 
their house to be built in Narcoossee. Once settled, they planted an 
orange grove. The letters trace their life through the birth of their 
daughter and Bill's job as a steamer captain. Employment troubles 
dogged the family and he attempted to sell the orange grove in June 
1887. When the letters come to an end, the family had not yet made 
any plans on what to do next. 

Many of the letters suffered from poor paper quality and bleed 
through from both sides of the pages causing the scanned cursive text 
to be hard to read. Time consuming and painstaking detailed image 
editing was conducted to enhance the natural lines of the text while 
omitting letters fading through from the opposite side. Once 
acceptable image quality had been achieved, the next step was to make 
the content of the letters accessible by a full text search. Since the 
letters were handwritten, no OCR software was available and limited 
full text options to uploading text transcripts to a large free text 
searchable field. Luckily, CONTENTdm provided such a field and the 
History Center had hard copy typed transcripts available with the 
letters. Lacking an electronic version of the transcripts, the transcripts 



Central Florida Memory 463 



were rekeyed and saved as individual text files appropriate for 
uploading with the images. 

Outsourcing 

The digitization of certain newspapers and college catalogs 
necessitated the outsourcing of certain functions such as microfilm 
scanning and optical character recognition. Lacking the original hard 
copies normally preferred for scanning, the microfilm versions of the 
Rollins College catalogs, St Cloud Tribune, and Lochmede newspaper 
were deemed to be of good enough quality for digitization. In addition 
to microfilm scanning, optical word recognition for some previously 
scanned newspapers and monographs, such as city and telephone 
directories, was outsourced as well. Several vendors were contacted 
and iArchives was selected to perform the services. With a proven 
history of creating high quality images and "highly accurate and 
patented OCR software which consistently produces higher accuracy 
results when compared to the 'off the shelf OCR engines," iArchives 
stood out from the rest (iArchives, 2010). During early discussions 
about metadata and file naming conventions, it was quickly realized 
that normal outputs differed slightly from the outsourced deliverables. 
Several test batches of materials were processed to give a better 
understanding of what could be expected. While some 
accommodations could be made, other aspects were fixed components 
of the iArchives system. Additional processing was necessary to bridge 
the gap between what the vendor could provide and what the 
expectations were for adding material to Central Florida Memory. 
However, one of the most basic differences was one that was not 
anticipated. The vendor's final images were delivered in a compressed 
format. While it did not affect the quality of the images, software had 
to be acquired and extra time was required to extract the individual 
files. 

During the IMLS grant phase, over 50,000 images were added to 
the collection and the website was redesigned. Focusing on creating a 
single collection rather than multiple collections defined by 
institution, the site itself began to rely more heavily on the contents of 
the collection. The first web design for CFM featured pages that were 



464 Central Florida Memory 

essentially a series of linked pictures that gave access to CONTENTdm 
and could not be easily updated. That design was improved upon in 
2004 when the pages were integrated with CONTENTdm. This gave 
the website a more cohesive look and made changes to its content 
easier. Unfortunately, it was found that upgrading CONTENTdm 
usually "broke" the design and adjustments and testing needed to be 
done each time. 

To cap off the grant year, a workshop was held for partners, 
potential partners, and others working in the digital collection field. 
Presentations focused on what had been and could be accomplished in 
building digital collections. The mission and goals of CFM were 
outlined and attention paid to how the project activities were 
organized and carried out. By providing examples of materials added, 
equipment and software acquired, and lesson plans created, 
institutions were shown how they could start their own projects. 

Work continued on the project after grant funding ended. Orange 
County Library System designed and implemented a CFM presence in 
Second Life while the UCF Libraries maintained and upgraded 
CONTENTdm and continued scanning activities. 

LSTA Grant - 2008 

In 2008, we applied for, and received, an LSTA grant to expand the 
collection by 30,000 images and our geographical coverage to include 
Volusia County. Similar to the previously funded IMLS grant, the 
LSTA grant was a good match for CFM as it was "a state-based 
program with a broad mandate to use technology to bring information 
to people in innovative and effective ways" (Library services, 2008). 
Bethune-Cookman University in Daytona Beach, Florida and Stetson 
University in DeLand, Florida were invited to become partners and to 
contribute material. Bethune-Cookman's contribution of historically 
significant images to CFM would increase the representation of early 
Central Florida's African-American community. Stetson's contribution 
of images and materials expanded CFM's geographic coverage beyond 
Orange and Seminole counties. To manage new contributions and an 
increased variety of materials, the Digital Projects Management 



Central Florida Memory 465 



System was updated. As before, the new system tracked materials 
through the use of uniform bibliographic information and a computer- 
generated unique identification number. Once an item was received, 
the new Digital Projects Management System then provided the 
means to monitor the item's progress by tracking staff input in the 
form of checklists and text entries at various stages of the production 
process. From monitoring data and statistics to exporting metadata, 
the system enabled skillful handling of a large number of unique 
requests. 

During the LSTA grant, the UCF Libraries' Digital Services unit 
was responsible for scanning and processing over 28,000 images. 
Previous experiences taught flexibility regarding production 
procedures and to prepare for the unexpected. Images were routinely 
reviewed for quality control and adjustments made to equipment and 
lighting. Due to the complex nature of scanning photographic and 
visual materials, it was decided that multiple equipment options 
would be employed to guarantee the highest quality images while 
protecting the original materials. By the end of the grant, four 
different types of scanning equipment had been necessary to carefully 
and accurately capture each image. Kodak i200 flatbed scanners were 
used to scan photographs in good condition and loosely bound 
materials. A newly purchased CopiBook overhead scanner proved 
effective for dealing with large photographs, yearbooks, catalogs, and 
unbound newsprint. Oversized and highly reflective materials were 
scanned on the Kaiser rePro graphic stand with Karden-Linhof 
camera. Tightly bound and moderately reflective items, such as bound 
glossy newspapers, were scanned on the Avision FB6080E A3 
Bookedge flatbed scanner. 

Once materials were scanned and images corrected, metadata was 
created by reviewing both the physical item and the scanned images. 
Since efficiencies in uploading small compound objects (one or two 
images) were achieved early on by using tab-delimited text files to 
upload preliminary metadata, this method was applied to larger 
compound objects as well. Controlled vocabulary fields were applied 
to 18 fields including author, photographer, subject, subject-personal 
names, publisher, studio, format, format-medium, subject-topic, 



466 Central Florida Memory 

repository, repository collection, type, language, coverage-spatial, 
coverage-temporal, funding source, date contributed, and CFM 
storyline. An important benefit of using controlled vocabulary fields is 
that they allowed the end user to perform a new search by linking 
metadata as phrases rather than individual words. In addition to 
making the images available to the public, master images were 
submitted to the Florida Digital Archive, "a cost-effective, long-term 
preservation repository for digital materials in support of teaching and 
learning, scholarship, and research in the state of Florida" (Florida 
Digital Archive, 2010). Once images were uploaded to Central Florida 
Memory, metadata records were exported and coupled with the 
respective images to create packages to be sent to the Florida Digital 
Archive for long term digital preservation. 




■ umt ten 10 m ■ 

Figure CFM-5: Central Florida Memory Website, 2009 (Central Florida Memory, 
2009) 



The website was completely redesigned using a different color 
palette and employing a cleaner look. CONTENTdm displays were 
pulled into the web pages but kept separate from them. The site could 



Central Florida Memory 467 



be updated and content changed easily and not be affected by any 
upgrades to the CONTENTdm software. 

The Future 

If the partners of Central Florida Memory have learned anything over 
the last seven years it is that whether or not funding is awarded, the 
project goes on. In funding years an extraordinary amount of work has 
been accomplished, and equipment and resources acquired. During 
the lean years when there is no funding, the work continues on a much 
smaller scale, but it goes on. The partners of CFM continue to meet 
quarterly, formulate new ideas on how to build the collection, improve 
access to the collection, and spread the message that this resource 
exists. New funding sources are being identified, new grants will be 
written, and new ideas will be implemented. The partners of CFM 
continue to believe in their commitment to the project and to each 
other. 

References 

Book 24: Carey Hand Funeral Home records, April 1, 1924 - August 
21, 1924. (n.d.). Central Florida Memory. Retrieved August 14, 
2005, from http://www.cfmemory.org 

Central Florida Memory grant writing committee. (2004). Central 
Florida Memory. [Institute of Museum and Library Services grant 
proposal]. Orlando, FL. Retrieved December 1, 2009, from 
http://www.cfmemory.org/Share/About/Narrative.pdf 

Central Florida Memory website. (2003). Retrieved May 11, 2003, 
from http://centralfloridamemory.lib.ucf.edu. 

Central Florida Memory website. (2005). Retrieved September 20, 
2005, from http://www.cfmemory.org 

Central Florida Memory website. (2009). Retrieved March 13, 2010, 
from http://www.cfmemory.org 

Digitization Spec Kit. (2005). Retrieved March 20, 2010, from http: 
//www.cfmemory.org/Share/Workshop/DigitizationSpecKit.pdf 



468 Central Florida Memory 

Digital Services Tracking Database. (2005). Internal tracking 

database used at University of Central Florida Libraries. Orlando, 
Florida. 

Dotson, L. (2005). See how it's done. Central Florida Workshop for 
Museum & Library Personnel. Retrieved March 5, 2010, from 
http : / / www. cfmemory. org/ Share /Workshop/Behind%20the%2oS 
cenes%20l.ppt 

Dunlop, D. (2005). How it all connects II. Central Florida Workshop 
for Museum & Library Personnel. Retrieved March 19, 2010, from 
http : / / www. cfmemory. org/Shar e /Workshop / 
Metadata%2oPart%202.ppt 

Florida digital archive. (2010). Retrieved March 8, 2010, from 
http:/ /www.fcla.edu/ digitalAr chive/. 

Gronlund, G. (2005). Weaving the story. Central Florida Workshop 
for Museum & Library Personnel. Retrieved March 27, 2010, from 
http : / / www. cfmemory. org/Shar e /Workshop / CENTRAL96 20 
FL0RIDA%20MEM0RY%20WEAVING%20THE%20ST0RY%209 

%2030%202005%202.ppt 

iArchives process. (2010). Retrieved February 20, 2010, from 
http : / / www.iarchives. com/process, shtml 

IMLS National Leadership grants. (2010). Retrieved February 17, 
2010, from http://www.imls.gov/applicants/grants/ 
NationalLeadership.shtm. 

Lavoie, J. (2005). Showing it to the world. Central Florida workshop 
for museum & library personnel. Retrieved February 23, 2010, 
from http://www.cfmemory.org/Share/Workshop/ 
Showing%2oit%20to%20the%2oWorld.ppt 

Library services and technology act grants. (2008). Retrieved 
January 14, 2010, from 

http://dlis.dos.state.fl.us/bld/grants/LSTA/ LSTAinfo.pdf 

Moving theory into practice digital imaging tutorial. (2010). 
Retrieved March 15, 2010, from 
http : / / www.library.cornell.edu / preservation/tutorial 



Central Florida Memory 469 



Orange County voter registration, (n.d.). Central Florida Memory. 
Retrieved August 15, 2005, from http://www.cfmemory.org 

PALMM documentation. (2010). Retrieved March 11, 2010, from 
http : / / fclaweb.fcla.edu /node/572/ 

Success story: Central Florida Memory. (2009). In A new story 
based approach to digital collection [OCLC CONTENTdm 
brochure]. Retrieved March 5, 2010, from 
http : / / www. oclc.org/ us/en/ services /brochures / 
211595usc_F_CentralFloridaMemory.pdf 



Apollo 13.0: Digitizing Astronaut Jack 
Swigert's Apollo Documents 



Andrew Weiss (Fort Hays State University, Forsyth Library) 



Abstract: 

The Kansas Cosmosphere and Space Center and Fort Hays State 
University Forsyth Library joined together in 2008 to digitize the 
personal archive of Apollo 13 astronaut John L. (Jack) Swigert. The 
documents (1966-1970) include blueprints and proofs of the Apollo 
Operations handbook Malfunction Procedures, which Swigert was 
called upon to revise in the aftermath of the Apollo 1 disaster in early 
1967. The partnership between the KCSC and Forsyth Library has 
been fruitful in the area of public relations and dissemination of 
digital collections online. Some of the obstacles related to the 
partnership have yielded positive results. As a result of the 
partnership we created a standard for uniform file names for Forsyth's 
Digital Collections, in adopting and customizing Goddard Core, a 
variation of Dublin Core, and in developing an External Partnership 
Protocol to improve future collaborative projects. 

Keywords: Apollo Operations, Best practices, Collaboration, 
CONTENTdm, Digitization, Dublin Core, Goddard Core, Handbook, 
Malfunction procedures, Museum-library partnership, Space 
exploration. 



470 



Apollo 13.0 471 



Introduction 

We shall not cease from exploration 
And the end of all our exploring 
Will be to arrive where we started 
And know the place for the first time. 

--T.S. Eliot, Four Quartets 

While orbiting the moon in 1968, Apollo 8 astronaut Bill Anders 
photographed Earthrise, one of the iconic images of the Apollo 
missions. Fragile and tiny, the earth appears "almost as a disk" in a 
dark void and concretely shows that humans are more miniscule than 
imagined (Brooks, Grimwood & Swenson, 1979, p. 277). Although 
later missions would overshadow Apollo 8, including the successful 
moon-landing of Apollo 11 and the successful failure of Apollo 13, the 
lasting legacy of this mission was a sudden appreciation for the Earth 
itself (Brooks et al., 1979, p. 366). The irony that a photograph of the 
Earth would become one of the most lasting images of the Apollo 
Program was not lost on the astronauts. Anders later remarked, "We 
came all this way to explore the moon, and the most important thing 
is that we discovered the earth" (Dordain, 2009, para. 6). 

This image, later adopted by environmentalists, eventually 
contributed to the general Zeitgeist that the mundane should take 
precedence over impractical dreams. Indeed, once an American had 
walked on the moon— as much a display of American real-politik 
might as a display of American ingenuity— public and political interest 
in the Apollo Program waned. By 1973 its funding was on its last legs 
and never again would the Apollo Program send an astronaut to the 
moon (McKie, 2008) (Brooks et al., 1979, p. 366). Yet if media 
coverage is any indication of an issue's timeliness, the 40th 
anniversary of the successful moon landing in 2009 proves that the 
spirit of the time remains vital. General interest in space exploration 
remains high among specialists and the general public alike. 
Furthermore, the improvement of digital technology is allowing us 



472 Apollo 13.0 



unprecedented access to the primary materials of important historical 
events, making the discoveries of the time even more immediate as 
they appear online. In essence, this is an exciting time to revisit the 
solid-state era of the Space Race of the 1950s, '60s and '70s through 
the power of digital technology. 

In its partnership with the Kansas Cosmosphere and Space 
Center (KCSC) to digitize the papers of John L. (Jack) Swigert, Fort 
Hays State University's Forsyth Library has had the privilege to work 
with documents of great importance to American and world history. 
At the same time we find ourselves looking inward and realize that an 
emphasis on the "mundane" still proves to be important. As a result of 
our work, we realize the need for a strong foundation in the 
development of sound policies and best practices that can work in 
reality. As the Apollo astronauts discovered, the point of a journey is 
not always to arrive; it's to find out where you've been. 

Background 

The Kansas Cosmosphere & Space Center (KCSC): 

Second only to the Smithsonian's Air and Space Museum in terms 
of collection size, the Kansas Cosmosphere and Space Center (KCSC), 
located in Hutchinson, Kansas, is one of the leading collectors of space 
artifacts in the United States; the items collected, many of which have 
flown in space, provide a complete overview of rocketry and space 
history from its origins in Nazi-era Germany, to the Space Race of the 
1950s and 1960s, and to the era of detente between Russia and 
America in the 1970s. Items housed include complete German V-i and 
V-2 rockets, a Redstone Nuclear Warhead, a Titan rocket, a Russian 
Vostok Spacecraft, The Mercury 7 Liberty Bell spacecraft, and the 
Apollo 13 Command Module. Their holdings also include the largest 
collection of Russian Space Program artifacts outside of the former 
Soviet Union (Kansas Cosmosphere, 2009). 

FHSU, Forsyth Library & FLDCI 

Forsyth Library is a small academic library serving the needs of 
Fort Hays State University's 11,000 full- and part-time students and 



Apollo 13.0 473 



275 faculty & staff members. Roughly 4,000 of those students are 
traditional on-campus undergraduate and graduate students. The 
other 7,000 students study through FHSU's Virtual College, an online 
degree program. The library is central to the Virtual College and 
provides assistance to those students in the form of a dynamic web 
presence that facilitates off-campus access to our OPAC, databases, 
and digital collections. This digital content includes the digitized 
Swigert documents, our Master's Thesis Collection, letters from 
Benjamin Franklin and historic glass-plate negatives collections. Our 
first collection of materials was digitized in 2004. The Forsyth Library 
Digital Collections Initiative (FLDCI) was established in 2008 in order 
to facilitate the development and preservation of digital content. 
Under the development of this framework of guidelines, we have been 
able to create full-fledged digitization partnerships both within our 
university, including the Sternberg Museum, and Graduate School, 
and externally with institutions in Kansas. Our modest digital 
collections are positioned to grow exponentially in the next few years, 
and the establishment of the FLDCI promises to simplify the growth 
process. 

Partnership Details 

The partnership between the Kansas Cosmosphere and Space 
Center and Forsyth Library began tentatively in 2003 with the 
proposal made by library director John Ross to digitize materials held 
by the KCSC. The primary goal for this first joint partnership was to 
scan, catalog and upload images from the Mercury Project Missions 
into an online content management system. The digitization team 
included the KCSC archivist, Kiersten Latham, Forsyth Library 
cataloger, Jerry Wilson, and Forsyth Library Archivist, Patty Nicholas. 
By the end of November 2004 the majority of images, which included 
photographs of the Mercury spacecraft, astronauts, and engineers, 
had been scanned. 

Early enhancements to the project began with the use of Dublin 
Core, which would later influence the presently used scheme. The 
second development was the decision to purchase CONTENTdm as 
the vehicle for the online dissemination of digital content. 



474 Apollo 13.0 

CONTENTdm was eventually purchased and installed for use in May 
2005. However, personnel changes at the KCSC and time limitations 
for Forsyth Library staff impacted both the development of image 
metadata and digitization. Consequently, no digital images from the 
Mercury Project were placed online until the project was revisited in 
March, 2010. 

In early 2008, in consultation with the new president and CEO of 
the KCSC, Chris Orwoll, and the new Collections Manager, Meredith 
Miller, another partnership was proposed by Forsyth Library director, 
John Ross, with the aim to digitize the archive of Apollo 13 astronaut 
John L. (Jack) Swigert. It was Swigert who had helped to revise the 
Apollo Operations Handbook Malfunction Procedures after the Apollo 
1 disaster of January 27, 1967. This collection of 10 archival boxes, 
each dedicated to a single spacecraft system, included unique 
unpublished materials, and represented an opportunity to refuel our 
partnership. As soon as copyright clearance was secured, we were 
given the green light to proceed. 

During the three years that elapsed between the two projects, 
CONTENTdm had been successfully implemented by Forsyth Library 
Systems Technician, Heath Bogart, and was being used for Forsyth 
Library Archives collections. Two digitization team members had been 
added to improve the organization and workflows of the suddenly 
burgeoning digitization projects, including new Catalog er and Digital 
Content Specialist, Sherry Severson, and new Digital Collections 
Librarian, Andrew Weiss. From August 2008 through May 2010, the 
entire collection of materials in Swigert's archive was digitized, and 
95% of these items were fully cataloged using Goddard Core, an 
elaboration of the Dublin Core metadata scheme developed by NASA's 
Goddard Museum. 

Digitization Goals 

This collection will help researchers and lay-people alike to see 
the program as a work-in-progress, something which can get lost as 
the Apollo mission becomes cemented in history. Our wider goal is to 
develop a collaboration model that will establish a set of procedures to 
use between a small academic library and a major museum. We 



Apollo 13.0 475 



believe this will eventually contribute to the foundation of a larger 
digitization project involving all phases of national and international 
space exploration extending from World War II to the mid-1970s. 

Apollo Mission Background: 

If we die, we want people to accept it. We're in a risky business, 
and we hope that if anything happens to us it will not delay the 
program. The conquest of space is worth the risk of life. 

- Gus Grissom (Barbour, 1969, p. 125) 

Along with John F. Kennedy's May 25, 1961 speech committing 
the United States to the "Space Race," arguably the most important 
date in the Apollo Program occurred on January 27th, 1967. On that 
tragic day, Virgil "Gus" Grissom, Ed White, and Roger Chaffee, 
became the first astronauts to die in the American Space Program. 
Their Command Module suddenly caught fire while the three were 
inside conducting a "Plugs Out" launch pad test one month before 
their scheduled mission (Bilstein, 1996, p. 340). Because of design 
flaws, the hatch could not be opened by the astronauts or blown open 
externally. The shocking part of the accident was that it occurred 
where none was expected: on Earth. Though a terrible disaster, the 
Apollo 1 accident fundamentally changed the Apollo Program and, in 
essence, prevented further accidents from happening while in space 
(Brooks et al., 1979, p. 214-225). 

After the accident of January 27th, a full investigation was 
conducted in order to arrive at the cause of the accident. During the 
investigation each piece of the Command Module was removed, using 
an identical copy of the Command Service Module (#014) as a guide, 
the engineers at NASA and contractor North American Aviation, Inc. 
(NAA) narrowed the cause down to two main factors. The first factor: 
the tests were being done in excessively high oxygen pressure, 
increasing the potential for fire. The second factor: too much Velcro 
was being used to tie wiring together. The combination of these two 
factors along with the spark from a faulty wire and the flaw in the 
hatch design, ironically implemented as a safety precaution, was 
ascertained as the cause of the accident (Orloff, 2004, para. 61) 
(Brooks et al., 1979, p. 214-225). 



476 Apollo 13.0 



The investigation marked a turning point in the development of 
the Apollo Program. In the early 1960s, when the mission to the moon 
became a national priority, NAA and NASA had agreed to a two-tiered 
project. Tier one would develop a Block I spacecraft (1962—1968), 
which was to be used for unmanned boilerplate missions and sub- 
orbital rocket testing, and tier 2 would develop a Block II spacecraft 
(1964—1973), which was to be used for manned earth- and lunar- 
orbital flights and the moon-landing missions (Brooks et al., 1979, p. 
229). The two phases overlapped between the years 1964 through 
1968, but once the Apollo 1 accident occurred, Block I was phased out, 
being used only for unmanned Apollo 4, 5 and 6 missions until early 
1968 (Brooks et al., 1979, p. 232). In order to redesign Block II, both 
engineers and astronauts were called upon to completely review the 
blueprints, diagrams and procedures for every aspect of the Apollo 
spacecraft. Included among these specialists was Astronaut John L. 
(Jack) Swigert, who reviewed the Apollo Spacecraft Malfunction 
Proc